<!DOCTYPE html>
<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body>
    P.S. Juergen mentioned the steep learning curve for the phylognetic
    comparative method. I am of course always happy to collaborate on
    specific projects - just drop me an email.<br>
    <br>
    G.<br>
    <br>
    <div class="moz-cite-prefix">On 11/6/23 08:23, Gerhard Jäger wrote:<br>
    </div>
    <blockquote type="cite"
      cite="mid:b4d009bf-0ede-4105-88a3-55b0b7f88e21@uni-tuebingen.de">
      <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
      Dear all,<br>
      <br>
      Thanks to Juegen Bohnemeyer for this thoughtful contribution. <br>
      <br>
      A couple of remarks.<br>
      <br>
      About the phylogenies we used in Jäger & Wahle 2021: We
      actually did not use the world tree. The reason for this choice
      was the fact that branch lengths are an important aspect of the
      phylogenetic comparative method. Branch lengths above the family
      level in the world tree (the one from <a
        class="moz-txt-link-freetext"
        href="https://www.nature.com/articles/sdata2018189"
        moz-do-not-send="true">https://www.nature.com/articles/sdata2018189</a>)
      are so obviously too short that this seriously undermines the
      method. (Btw, the phylogenetic comparative method is fairly robust
      regarding errors in the topology of the phylogenies, but branch
      lengths are important). <br>
      <br>
      Instead of using the world tree, we inferred a posterior sample of
      phylogenies for each Glottolog family. So like Dunn et al., we
      used separate trees for different families. The main technical
      innovation - besides the fact that we used dozens of families
      instead of four - was that we tested using the same Markov Chain
      parameters across families, where Dunn et al. fitted a separate
      model for each family. Our's is effectively an implementation of
      Maslova's ideas.<br>
      <br>
      In the extreme case where each family contains just one language
      in a sample, the Jäger & Wahle method is actually equivalent
      to stratified sampling where one language is sampled from each
      family. So our method is not so much an alternative to stratified
      sampling but an extension that allows to use all languages for
      which you have data.<br>
      <br>
      Regarding the phylogenies derived from ASJP: Juergen, you say,
      "Notably, Jäger’s method takes into account systematic sound
      correspondences, so to my uninitiated eye, it looks like an
      unassailable computational implementation of the
      historic-comparative method." I am afraid you are doing me too
      much honor here. My method takes recurring sound correspondences
      into account, but it is still a far cry from a full implementation
      of the historic-comparative method. However, in Jäger 2018 (<a
        class="moz-txt-link-freetext"
        href="https://www.nature.com/articles/sdata2018189"
        moz-do-not-send="true">https://www.nature.com/articles/sdata2018189</a>),
      I ran a lot of tests demonstrating that the quality of the
      phylogenies obtained this way is fairly adequate (using Glottolog
      as gold standard) despite the method's conceptual shortcomings.<br>
      <br>
      I agree with Juergen that there is no consensus approach for the
      phylogenetic comparative method. I am not sure though that such a
      consensus is possible or even desirable. A linguistic result -
      say, the universal correlation between adposition-noun order and
      verb-object order - is trustworthy if several different methods
      confirm it. These methods may include stratified sampling with
      different sampling protocols, family bias estimation, phylogenetic
      comparative methods with different trees and different priors,
      perhaps others. If some approaches confirm it and others do not,
      that is a reason to dig deeper.<br>
      <br>
      Best, Gerhard<br>
      <br>
      <br>
      <br>
      <div class="moz-cite-prefix">On 11/6/23 00:22, Juergen Bohnemeyer
        wrote:<br>
      </div>
      <blockquote type="cite"
cite="mid:SJ0PR15MB46966D0B2BE32910582F1528DDA4A@SJ0PR15MB4696.namprd15.prod.outlook.com">
        <meta http-equiv="Content-Type"
          content="text/html; charset=UTF-8">
        <meta name="Generator"
          content="Microsoft Word 15 (filtered medium)">
        <style>@font-face
        {font-family:Helvetica;
        panose-1:0 0 0 0 0 0 0 0 0 0;}@font-face
        {font-family:Wingdings;
        panose-1:5 0 0 0 0 0 0 0 0 0;}@font-face
        {font-family:SimSun;
        panose-1:2 1 6 0 3 1 1 1 1 1;}@font-face
        {font-family:"Cambria Math";
        panose-1:2 4 5 3 5 4 6 3 2 4;}@font-face
        {font-family:Calibri;
        panose-1:2 15 5 2 2 2 4 3 2 4;}@font-face
        {font-family:"CMU Serif";
        panose-1:2 0 6 3 0 0 0 0 0 0;}@font-face
        {font-family:"Times New Roman \(Body CS\)";
        panose-1:2 11 6 4 2 2 2 2 2 4;}@font-face
        {font-family:"\@SimSun";
        panose-1:2 1 6 0 3 1 1 1 1 1;}p.MsoNormal, li.MsoNormal, div.MsoNormal
        {margin:0in;
        font-size:10.0pt;
        font-family:"Calibri",sans-serif;}a:link, span.MsoHyperlink
        {mso-style-priority:99;
        color:blue;
        text-decoration:underline;}p.MsoListParagraph, li.MsoListParagraph, div.MsoListParagraph
        {mso-style-priority:34;
        margin-top:0in;
        margin-right:0in;
        margin-bottom:0in;
        margin-left:.5in;
        font-size:10.0pt;
        font-family:"Calibri",sans-serif;}span.EmailStyle19
        {mso-style-type:personal-reply;
        font-family:"CMU Serif";
        color:windowtext;
        font-weight:normal;
        font-style:normal;}.MsoChpDefault
        {mso-style-type:export-only;
        font-size:10.0pt;
        mso-ligatures:none;}div.WordSection1
        {page:WordSection1;}ol
        {margin-bottom:0in;}ul
        {margin-bottom:0in;}</style>
        <div class="WordSection1">
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif"">Dear
              all – Two points:<o:p></o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif"">First,
              this thread has served to alleviate somewhat my concerns
              about the review processes used by the generic science
              journals – mostly through offline one-on-one
              communications I’ve received. Thanks!<o:p></o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif"">Secondly,
              I’d like to, maybe gently push back, or may be just
              elaborate, on the contrast Simon Greenhill draws between
              phylogenetic inference and the “overly simplistic
              correlation methods”, as he calls them. In the process,
              I’d like to try to clarify what in my mind are the
              strengths and weaknesses of each (type of) method. And
              I’ll pose a couple of challenge questions to the
              detractors of phylogenetic/coevolutionary testing and one
              for their proponents, and I hope all of these questions
              won’t go unanswered </span> <span
style="font-size:12.0pt;font-family:"Apple Color Emoji"">😊</span><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p></o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif"">Before
              I do that, let me flag my interest in this: I’m currently
              in Leipzig spending a sabbatical with Russell Gray’s
              Department of Linguistic and Cultural Evolution (</span><span
style="font-size:12.0pt;font-family:"Cambria Math",serif;mso-fareast-language:ZH-CN">≈
            </span><span
              style="font-size:12.0pt;font-family:"CMU Serif"">the
              Grambank team), with the principal objective of figuring
              out how to capitalize the most on Grambank for my project.
              The project tracks the typological distribution of
              functional expressions and attempts to account for it (the
              distribution) in terms of an evolutionary process that
              makes grammaticalization sensitive to functional
              pressures. Since coming here, I’ve been struggling with
              the question whether I should adopt stratified sampling, a
              phylogenetic method (I’m not going to use the term
              ‘co-evolution’ here, because I’m at this stage of the
              project just looking at the emergence of individual
              functional expressions), or both. So I’m airing my
              thoughts very much in hopes of eliciting feedback.</span><span
style="font-size:12.0pt;font-family:"Cambria Math",serif;mso-fareast-language:ZH-CN"><o:p></o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif"">So.
              Let’s start with a very basic reminder: We cannot observe
              causality, we can only infer it, as David Hume taught us
              (i.e., humanity) nearly three centuries ago. The
              difference between the methods at issue is one between
              synchronic inference and diachronic inference. <o:p></o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif"">Now,
              before I go any further, a big disclaimer: I think I
              understand at the most abstract level the ideas underlying
              phylogenetic inference. But I don’t understand the
              statistical “sausage making” involved. Meaning I have many
              basic questions about the algorithms used for phylogenetic
              inference. <o:p></o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif"">With
              that said: Imagine an ideal world in which we had perfect
              and reliable knowledge about the complete phylogeny of
              every extant human language and every language that ever
              existed. It seems indisputable to me that in this world,
              we would test hypotheses about causal relations between
              two linguistic features not merely by looking at their
              synchronic distributions, but also by looking at whether
              the two features tend to co-evolve across phylogenies or
              whether they appear to have merely accidentally “travelled
              together” through time in a few families. <o:p></o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif"">So
              this would be my first question to the detractors of
              phylogenetic/dynamic (Maslova’s term)/co-evolutionary
              tests for typological generalizations: do you disagree
              that if we had that perfect knowledge, we would as a
              matter of course take into account the phylogenetic
              perspective?<o:p></o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif"">Now,
              in the real world, our knowledge of phylogenies is itself
              largely inferred from synchronic distributions. Crucially,
              though, these inferences are based on data that may
              overlap with the typological patterns of interest, but is
              largely independent of it. To make this more concrete:<o:p></o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <ul style="margin-top:0in" type="disc">
            <li class="MsoListParagraph"
              style="margin-left:0in;mso-list:l0 level1 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">Jäger &
                Wahle (2021) use the world tree inferred from <a
                  href="https://asjp.clld.org/" moz-do-not-send="true">ASJP</a>
                cognate data presented in Jäger (2018). Notably, Jäger’s
                method takes into account systematic sound
                correspondences, so to my uninitiated eye, it looks like
                an unassailable computational implementation of the
                historic-comparative method. Of course, the data the
                analysis is based on is another matter.<o:p></o:p></span></li>
          </ul>
          <p class="MsoListParagraph"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <ul style="margin-top:0in" type="disc">
            <li class="MsoListParagraph"
              style="margin-left:0in;mso-list:l0 level1 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">Verkerk et
                al. (ms.), which Martin cites in the post that started
                the thread, rely on Glottolog, which Jäger & Wahle
                also used to validate their inferred world tree. In my
                understanding, Glottolog is sourced by a compilation of
                the best available evidence for phylogenetic relations,
                principally curated by its lead author, Harald
                Hammarström.<o:p></o:p></span></li>
          </ul>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif"">Two
              more points to keep in mind here:<o:p></o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <ul style="margin-top:0in" type="disc">
            <li class="MsoListParagraph"
              style="margin-left:0in;mso-list:l0 level1 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">Stratified
                sampling of course also takes into account phylogenetic
                and areal information. So the two types of approaches
                have really more in common than Simon gives them credit
                for in his reply, I think. However, stratified sampling
                uses phylogenetic and areal information in first
                approximation only “negatively”, as it were: by
                systematically removing observations from the analysis
                that are suspected to be “tainted” by areal or
                genealogical dependencies. <o:p></o:p></span></li>
          </ul>
          <p class="MsoListParagraph"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <ul style="margin-top:0in" type="disc">
            <li class="MsoListParagraph"
              style="margin-left:0in;mso-list:l0 level1 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">The distance
                of our real world from that ideal world in which we have
                perfect knowledge of phylogenetic and areal relations is
                in my (again, uninitiated, in the sense that I’m not a
                historical linguist) mind pretty darn great. However,
                the phylogenetic statistics algorithms use “forests” of
                alternative phylogenetic trees weighted for the
                confidence the field (or the analysts) has/have in them,
                factored into the tests as Bayesian priors. In this
                sense, the use of phylogenetic information is more
                sophisticated in the dynamic tests than in anything
                based on stratified sampling that I have seen.<o:p></o:p></span></li>
          </ul>
          <p class="MsoListParagraph"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif"">So
              with all this assembled, here’s my semi-informed take on
              the pros and cons of each approach:<o:p></o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <ul style="margin-top:0in" type="disc">
            <li class="MsoListParagraph"
              style="margin-left:0in;mso-list:l0 level1 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">Stratified
                sampling (I’m going to start with the method that has
                been practiced the longest)<o:p></o:p></span></li>
          </ul>
          <p class="MsoListParagraph"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <ul style="margin-top:0in" type="disc">
            <ul style="margin-top:0in" type="circle">
              <li class="MsoListParagraph"
                style="margin-left:0in;mso-list:l0 level2 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">Pro<o:p></o:p></span></li>
              <ul style="margin-top:0in" type="square">
                <li class="MsoListParagraph"
                  style="margin-left:0in;mso-list:l0 level3 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">Technologically
                    and conceptually simple, easy to implement by
                    anybody with the most basic level of training in
                    inferential statistics, and transparent for the
                    analyst to track and understand the effects of their
                    decisions.<o:p></o:p></span></li>
                <li class="MsoListParagraph"
                  style="margin-left:0in;mso-list:l0 level3 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">Statistically
                    conservative, in the specific sense that it
                    minimizes the odds of false positives, i.e., of
                    support for generalizations for which the evidence
                    isn’t really there. (One might think that Dunn et
                    al. 2011 present counterevidence to this. But the
                    apparent false positives pointed out by Dunn et al.
                    turn out to be true positives after all, now that
                    Jäger & Wahle and Verkerk et al. have replicated
                    their analysis with a larger sample.)<o:p></o:p></span></li>
                <li class="MsoListParagraph"
                  style="margin-left:0in;mso-list:l0 level3 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">Validity of
                    testing depends only on one key assumption – the big
                    one: independence of observations. (Detractors of
                    stratified sampling might however argue that it is
                    impossible to ensure that this assumption is
                    warranted in any decent-sized sample.)<o:p></o:p></span></li>
              </ul>
              <li class="MsoListParagraph"
                style="margin-left:0in;mso-list:l0 level2 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">Con<o:p></o:p></span></li>
              <ul style="margin-top:0in" type="square">
                <li class="MsoListParagraph"
                  style="margin-left:0in;mso-list:l0 level3 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">There has
                    never been a consensus on an optimal sampling
                    algorithm that balances minimization of areal and
                    genealogical biases with retaining enough
                    statistical power to detect the patterns of
                    interest. With such a consensus absent after 45
                    years of some of the smartest minds in the field
                    applying themselves to the problem, maybe there is
                    no optimal solution to be had?<o:p></o:p></span></li>
                <li class="MsoListParagraph"
                  style="margin-left:0in;mso-list:l0 level3 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">Based on
                    information reduction (exclusion of observations);
                    does not take into account phylogenetic information
                    to the fullest measure of what can be concluded from
                    it.<o:p></o:p></span></li>
              </ul>
            </ul>
          </ul>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <ul style="margin-top:0in" type="disc">
            <li class="MsoListParagraph"
              style="margin-left:0in;mso-list:l0 level1 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">Dynamic/phylogenetic/co-evolutionary
                inference<o:p></o:p></span></li>
          </ul>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <ul style="margin-top:0in" type="disc">
            <ul style="margin-top:0in" type="circle">
              <li class="MsoListParagraph"
                style="margin-left:0in;mso-list:l0 level2 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">Pro<o:p></o:p></span></li>
              <ul style="margin-top:0in" type="square">
                <li class="MsoListParagraph"
                  style="margin-left:0in;mso-list:l0 level3 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">Utilizes the
                    best available phylogenetic information “for all
                    it’s worth”, directly examining whether features
                    tend to coevolve across lineages or only in
                    particular lineages. (The same cannot currently be
                    said about areal information. However, Verkerk et
                    al. (ms.) use mixed models regression – which is
                    feasible based on Grambank thanks to its relatively
                    low percentage of missing observations – to factor
                    in possible areal effects.)<o:p></o:p></span></li>
                <li class="MsoListParagraph"
                  style="margin-left:0in;mso-list:l0 level3 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">Does not
                    depend on the assumption of independence of
                    observation.<o:p></o:p></span></li>
              </ul>
            </ul>
          </ul>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <ul style="margin-top:0in" type="disc">
            <ul style="margin-top:0in" type="circle">
              <li class="MsoListParagraph"
                style="margin-left:0in;mso-list:l0 level2 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">Con<o:p></o:p></span></li>
              <ul style="margin-top:0in" type="square">
                <li class="MsoListParagraph"
                  style="margin-left:0in;mso-list:l0 level3 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">Very steep
                    training demands both on the conceptual and on the
                    technological side. Currently only a handful of
                    specialists are able to perform such analyses. In
                    addition, the computational complexities are so
                    demanding as to render the analyses effectively
                    opaque for all but those specialists, raising
                    questions about the validation of such tests.
                    (Gerhard Jäger’s comparison to carbon dating in
                    archeology is apt here as well, I think: while
                    phylogenetic analysis will likely become more
                    standard in the future, it may be the case that
                    typologists will always have to rely on specialists
                    for actually performing such tests, just like
                    archeologists rely on physicists or lab technicians
                    for performing carbon dating. I think </span> <span
style="font-size:12.0pt;font-family:"Apple Color Emoji"">😉</span><span
style="font-size:12.0pt;font-family:"CMU Serif"">)<o:p></o:p></span></li>
                <li class="MsoListParagraph"
                  style="margin-left:0in;mso-list:l0 level3 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">Reliability
                    of phylogenetic inference seems to depend on the
                    size of the phylogenies involved: the smaller the
                    family, the higher the uncertainty in the absence of
                    actual historic data.<o:p></o:p></span></li>
                <li class="MsoListParagraph"
                  style="margin-left:0in;mso-list:l0 level3 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">As with
                    stratified sampling, there does not yet appear to be
                    a consensus approach. </span><span
style="font-size:12.0pt;font-family:"CMU Serif"" lang="DE">Jäger
                    & Wahle (2021) and Verkerk et al. </span><span
style="font-size:12.0pt;font-family:"CMU Serif"">(ms.) use
                    different algorithms (implemented in different
                    software packages), make different assumptions, and
                    their tests are conceptually quite distinct. <o:p></o:p></span></li>
                <li class="MsoListParagraph"
                  style="margin-left:0in;mso-list:l0 level3 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">Outcomes
                    depend on many assumptions which are surrounded by
                    varying degrees of uncertainty. <o:p></o:p></span></li>
              </ul>
            </ul>
          </ul>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif"">So
              at long last, here are my three challenge questions:<o:p></o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <ul style="margin-top:0in" type="disc">
            <li class="MsoListParagraph"
              style="margin-left:0in;mso-list:l0 level1 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">For the
                detractors of phylogenetic inference: <o:p></o:p></span></li>
            <ul style="margin-top:0in" type="circle">
              <li class="MsoListParagraph"
                style="margin-left:0in;mso-list:l0 level2 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">(Repeated
                  from above to make sure this doesn’t get lost) Do you
                  (dis)agree that if we had complete and robust
                  knowledge of the phylogenies of the extant languages,
                  we would perform phylogenetic tests as a matter of
                  course when attempting to validate typological
                  generalizations?<o:p></o:p></span></li>
              <li class="MsoListParagraph"
                style="margin-left:0in;mso-list:l0 level2 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">Do you
                  dispute that since phylogenetic testing of typological
                  generalizations has become possible, typology as a
                  field cannot simply ignore it and move on without it?
                  <o:p></o:p></span></li>
            </ul>
            <li class="MsoListParagraph"
              style="margin-left:0in;mso-list:l0 level1 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">For the
                proponents of phylogenetic inference: <o:p></o:p></span></li>
            <ul style="margin-top:0in" type="circle">
              <li class="MsoListParagraph"
                style="margin-left:0in;mso-list:l0 level2 lfo1"><span
style="font-size:12.0pt;font-family:"CMU Serif"">Do you
                  dispute that stratified sampling is statistically
                  conservative and has the advantage of transparency and
                  robustness vis-à-vis the underlying assumptions, so it
                  is likely not a method that typology will simply move
                  beyond and disregard in the future?<o:p></o:p></span></li>
            </ul>
          </ul>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif"">And
              there it. Apologies for the overlong post! – Cheers –
              Juergen<o:p></o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif""><o:p> </o:p></span></p>
          <div>
            <div>
              <p class="MsoNormal"><span
style="font-size:9.0pt;font-family:Helvetica;color:black">Juergen
                  Bohnemeyer (He/Him)<br>
                  Professor, Department of Linguistics<br>
                  University at Buffalo <br>
                  <br>
                  Office: 642 Baldy Hall, UB North Campus<br>
                  Mailing address: 609 Baldy Hall, Buffalo, NY 14260 <br>
                  Phone: (716) 645 0127 <br>
                  Fax: (716) 645 3825<br>
                  Email: </span><span style="font-size:11.0pt"><a
                    href="mailto:jb77@buffalo.edu"
                    title="mailto:jb77@buffalo.edu"
                    moz-do-not-send="true"><span
style="font-size:9.0pt;font-family:Helvetica;color:#0078D4">jb77@buffalo.edu</span></a></span><span
style="font-size:9.0pt;font-family:Helvetica;color:black"><br>
                  Web: </span><span style="font-size:11.0pt"><a
                    href="http://www.acsu.buffalo.edu/~jb77/"
                    title="http://www.acsu.buffalo.edu/~jb77/"
                    moz-do-not-send="true"><span
style="font-size:9.0pt;font-family:Helvetica;color:#0563C1">http://www.acsu.buffalo.edu/~jb77/</span></a></span><span
style="font-size:9.0pt;font-family:Helvetica;color:black"> <br>
                  <br>
                </span><span style="font-size:11.0pt;color:black">Office
                  hours Tu/Th 3:30-4:30pm in 642 Baldy or via Zoom
                  (Meeting ID 585 520 2411; Passcode Hoorheh) </span><span
style="font-size:9.0pt;font-family:Helvetica;color:black"><br>
                  <br>
                  There’s A Crack In Everything - That’s How The Light
                  Gets In <br>
                  (Leonard Cohen)  </span><span style="font-size:11.0pt"><o:p></o:p></span></p>
              <p class="MsoNormal"><span style="font-size:11.0pt">-- <o:p></o:p></span></p>
            </div>
          </div>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif""
              lang="DE"><o:p> </o:p></span></p>
          <p class="MsoNormal"><span
              style="font-size:12.0pt;font-family:"CMU Serif""
              lang="DE"><o:p> </o:p></span></p>
          <div id="mail-editor-reference-message-container">
            <div>
              <div
style="border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0in 0in 0in">
                <p class="MsoNormal"
style="mso-margin-top-alt:0in;margin-right:0in;margin-bottom:12.0pt;margin-left:.5in">
                  <b><span style="font-size:12.0pt;color:black">From: </span></b><span
                    style="font-size:12.0pt;color:black">Lingtyp <a
                      class="moz-txt-link-rfc2396E"
href="mailto:lingtyp-bounces@listserv.linguistlist.org"
                      moz-do-not-send="true"><lingtyp-bounces@listserv.linguistlist.org></a>
                    on behalf of Simon Greenhill <a
                      class="moz-txt-link-rfc2396E"
                      href="mailto:simon@simon.net.nz"
                      moz-do-not-send="true"><simon@simon.net.nz></a><br>
                    <b>Date: </b>Saturday, November 4, 2023 at 13:39<br>
                    <b>To: </b><a
class="moz-txt-link-abbreviated moz-txt-link-freetext"
                      href="mailto:lingtyp@listserv.linguistlist.org"
                      moz-do-not-send="true">lingtyp@listserv.linguistlist.org</a>
                    <a class="moz-txt-link-rfc2396E"
                      href="mailto:lingtyp@listserv.linguistlist.org"
                      moz-do-not-send="true"><lingtyp@listserv.linguistlist.org></a><br>
                    <b>Cc: </b>Dunn Michael <a
                      class="moz-txt-link-rfc2396E"
                      href="mailto:michael.dunn@lingfil.uu.se"
                      moz-do-not-send="true"><michael.dunn@lingfil.uu.se></a>,
                    Russell Gray <a class="moz-txt-link-rfc2396E"
                      href="mailto:russell_gray@eva.mpg.de"
                      moz-do-not-send="true"><russell_gray@eva.mpg.de></a><br>
                    <b>Subject: </b>[Lingtyp] Greenbergian word order
                    universals: confirmed after all<o:p></o:p></span></p>
              </div>
              <div>
                <p class="MsoNormal" style="margin-left:.5in"><span
                    style="font-size:11.0pt">Colleagues, Martin,
                    everyone else<br>
                    <br>
                    Thank you for sharing your perspectives on our 2011
                    paper. It's nice to see this still be discussed more
                    than a decade later. However, I would like to
                    express my concerns and disagreements with some of
                    the points you've raised.<br>
                    <br>
                    I'm very proud of the Dunn et al. paper for a number
                    of reasons. I'll name three.<br>
                    <br>
                    First, the paper showed that the overly simplistic
                    correlation methods that had been used to make
                    sweeping global claims were problematic. We need
                    better tools to tackle these questions, and the
                    tools we applied were one part of a better toolkit.
                    <br>
                    <br>
                    Second, it highlighted the need to understand
                    language systems in a diachronic manner. We cannot
                    decouple language typology from language history,
                    instead we need to understand how these are
                    entangled.<br>
                    <br>
                    Third, it emphasised the way that particular
                    configurations of languages can be arrived at via
                    different routes in different families at different
                    times. This enables a much richer understanding of
                    how these particular generalisations have arisen.<br>
                    <br>
                    Have Jäger and Wähle disproved any of that? no.
                    Maybe these were not completely novel insights
                    (Maslova’s work has been mentioned which touches on
                    a few of these issues too, for example), but these
                    ideas did appear to crystallise in this paper.<br>
                    <br>
                    While it's certainly important to revisit and
                    reevaluate research findings to ensure accuracy, it
                    is crucial to approach these discussions with an
                    understanding of the scientific process. Scientific
                    paradigms evolve over time, and different studies
                    may yield varying results due to changes in
                    methodologies, data sources, and sample sizes. This
                    doesn't necessarily imply that the initial research
                    was flawed or that the authors were neglectful. In
                    particular, the tools, the data, and our
                    understanding of how languages change are
                    substantially further advanced than they were a
                    decade ago (or, I know that *my* understanding of
                    these things is more advanced now, at least). And
                    these other papers that you mention -- and many
                    other studies -- have built upon the work we did in
                    2011. <br>
                    <br>
                    Furthermore, I would like to caution against drawing
                    overly broad conclusions about the quality of
                    research published in high-prestige journals. The
                    peer-review process in such journals is rigorous,
                    and while they may occasionally feature
                    sensationalist claims, this doesn't diminish the
                    overall value they contribute to the scientific
                    community. For the record, of the handful of papers
                    I've had in these journals *all* have been reviewed
                    by people I would infer to be linguists based on the
                    comments and issues they raised. We did not send
                    these papers to these journals to avoid linguistic
                    reviewers but, frankly, I've had better reviews at
                    these journals than at prominent linguistics
                    journals (and by "better" I mean more rigorous, more
                    thorough, and more critical).<br>
                    <br>
                    Finally, linguistic typology is an ongoing and
                    evolving field trying to tackle very difficult
                    problems. We need all the tools and approaches we
                    can get to solve these problems across all the
                    levels that languages operate on (from detailed
                    language internal analyses to high-level global
                    analyses). Rather than looking back and gate-keeping
                    what is 'real’ typology published in ‘real’
                    linguistics journals, we should shift our focus
                    forward. Typology can be a welcoming and diverse
                    community that embraces a wide range of approaches,
                    analyses, and styles. Let's look outward to foster
                    connections with other fields and disciplines. <br>
                    <br>
                    After all, why shouldn't linguistic typology work be
                    everywhere in science? it's certainly interesting
                    enough. <br>
                    <br>
                    Simon<br>
                    <br>
                    Dr. Simon J. Greenhill<br>
                    <br>
                    Associate Professor <br>
                    <br>
                    Te Kura Mātauranga Koiora | School of Biological
                    Sciences<br>
                    Te Whare Wānanga o Tāmaki Makaurau | University of
                    Auckland<br>
                    <br>
                    Abteilung für Sprach- und Kulturevolution |
                    Department of Linguistic and Cultural Evolution<br>
                    Max-Planck-Institut für Evolutionäre Anthropologie |
                    Max Planck Institute for Evolutionary Anthropology<br>
                    <br>
                    _______________________________________________<br>
                    Lingtyp mailing list<br>
                    <a
class="moz-txt-link-abbreviated moz-txt-link-freetext"
                      href="mailto:Lingtyp@listserv.linguistlist.org"
                      moz-do-not-send="true">Lingtyp@listserv.linguistlist.org</a><br>
                    <a
href="https://listserv.linguistlist.org/cgi-bin/mailman/listinfo/lingtyp"
                      moz-do-not-send="true">https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Flistserv.linguistlist.org%2Fcgi-bin%2Fmailman%2Flistinfo%2Flingtyp&data=05%7C01%7Cjb77%40buffalo.edu%7Cbe1181a9879e434ccc0308dbdd3319c4%7C96464a8af8ed40b199e25f6b50a20250%7C0%7C0%7C638346983806627973%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=OgjDO%2FGow%2FfFmxOIvmrWNNq1nwcsZaLzI5OMnoGjBXg%3D&reserved=0</a><o:p></o:p></span></p>
              </div>
            </div>
          </div>
        </div>
        <br>
        <fieldset class="moz-mime-attachment-header"></fieldset>
        <pre class="moz-quote-pre" wrap="">_______________________________________________
Lingtyp mailing list
<a class="moz-txt-link-abbreviated moz-txt-link-freetext"
        href="mailto:Lingtyp@listserv.linguistlist.org"
        moz-do-not-send="true">Lingtyp@listserv.linguistlist.org</a>
<a class="moz-txt-link-freetext"
href="https://listserv.linguistlist.org/cgi-bin/mailman/listinfo/lingtyp"
        moz-do-not-send="true">https://listserv.linguistlist.org/cgi-bin/mailman/listinfo/lingtyp</a>
</pre>
      </blockquote>
      <br>
      <pre class="moz-signature" cols="72">-- 
Prof. Dr. Gerhard Jäger
Universität Tübingen
Seminar für Sprachwissenschaft
Tel.: +49-7071-29-77302
<a class="moz-txt-link-freetext"
      href="http://www.sfs.uni-tuebingen.de/~gjaeger/"
      moz-do-not-send="true">http://www.sfs.uni-tuebingen.de/~gjaeger/</a></pre>
    </blockquote>
    <br>
    <pre class="moz-signature" cols="72">-- 
Prof. Dr. Gerhard Jäger
Universität Tübingen
Seminar für Sprachwissenschaft
Tel.: +49-7071-29-77302
<a class="moz-txt-link-freetext" href="http://www.sfs.uni-tuebingen.de/~gjaeger/">http://www.sfs.uni-tuebingen.de/~gjaeger/</a></pre>
  </body>
</html>