<!DOCTYPE html>
<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body>
    <p>I agree with Peter that the corpus-based methods employed by
      Hawkins, Wälchli, Cysouw, Levshina and others have been very
      important, and also with Jürgen that "when confronting the causal
      inference problem in typology, we must consider every source of
      evidence that we can get our hands on."</p>
    <p>But I don't agree with Peter that "the whole enterprise [of overcoming
      genealogical and areal biases] does not appear to be very
      productive", and I don't agree with Jürgen that we "must
      eventually move from secondary data typology to primary data
      typology".</p>
    <p>I think that the enterprise of controlling for family and contact
      effects is absolutely necessary, because otherwise we cannot
      distinguish outcomes of universal/non-historical factors from
      outcomes of historical events. Peter recognizes this implicitly
      when he says that we should "combine experimental research ...
      with a quantitative study of variation in corpora across a small
      number of sufficiently distinct languages". That's precisely the
      point: Which languages are "sufficiently distinct"? And hasn't the
      search for empirical universals been *highly productive* over the
      last few decades? The recent paper by Verkerk et al. (2025) has
      found good evidence for most of the empirical universals that had
      been seriously discussed earlier, so the Greenbergian universals
      seem to very robust findings compared to many other prestigious
      claims in linguistics.</p>
    <p>And I think that there is no reason to abandon secondary-data
      typology just because we can also (increasingly) do primary-data
      typology. Typological comparison can be done at multiple scales
      and multiple levels of granularity, and it is not clear that we
      can dispense with any of these levels. For example, we want to do
      typology of phonological segments (along the lines of the
      Phoible.org database), or typology of word meanings (lexification
      typology, cf. <a class="moz-txt-link-freetext" href="https://clics.clld.org/">https://clics.clld.org/</a>), and for these, it seems
      that secondary data will not be easily replaced.</p>
    <p>Best,</p>
    <p>Martin</p>
    <p><br>
    </p>
    <div class="moz-cite-prefix">On 21.11.25 16:04, Juergen Bohnemeyer
      wrote:<br>
    </div>
    <blockquote type="cite"
cite="mid:SJ0PR15MB46964946AEEB120D3A4BD26BDDD5A@SJ0PR15MB4696.namprd15.prod.outlook.com">
      <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
      <div dir="ltr"
style="font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
        Dear Peter — I’m a massive fan of corpus-based typology. More
        broadly, there is no question in my mind that we should, and
        must, eventually move from secondary data typology to primary
        data typology. Nobody seems to deny that secondary data typology
        is fraught with too many problematic idealizations: in
        particular, it reduces entire languages to single observations,
        and it suffers from incomparable decisions on what is treated as
        a language in different parts of the world. </div>
      <div dir="ltr"
style="font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
        <br>
      </div>
      <div dir="ltr"
style="font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
        (The second problem is closely related to, but not entirely
        identical with, the countability problem Ian Joo mentions. The
        fact that
        <i>language</i> is a count noun is a powerful illustration of
        how ordinary language can frame reality in ways that may impede
        scientific progress if it goes unchecked, as Whorf pointed out.
        However, actually counting languages is not the issue for
        regression-based modeling, since regression models don’t operate
        on counts. But the question whether what is treated as an
        observation (i.e., a language) is uniform across the sample is
        of course very much a concern for the validity of sampling-based
        and regression-based modeling alike.)</div>
      <div dir="ltr"
style="font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
        <br>
      </div>
      <div dir="ltr"
style="font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
        There is a broader answer to your question, though: as a matter
        of course, when confronting the causal inference problem in
        typology (i.e., when hunting for the causal forces that shape
        languages), we must consider every source of evidence that we
        can get our hands on.  Aside from corpus-based typology, this
        includes field-based psycholinguistics and the toolkit of
        evolutionary linguistics, including simulations and miniature
        artificial language experiments. </div>
      <div dir="ltr"
style="font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
        <br>
      </div>
      <div dir="ltr"
style="font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
        Let me also suggest a distinction between methods that are
        primarily geared toward the discovery of typological
        distributions and the examination of their statistical
        properties and methods than can be used to test hypotheses of
        causal inference (i.e., explanatory hypotheses). Experimental
        research such as what I just mentioned has its uses primarily
        for testing explanatory hypotheses. Corpus-based research can
        have both functions. But if we want to use corpora to discover
        typological distributions, we’ll need very large parallax corpus
        databases. As are being developed now. </div>
      <div dir="ltr"
style="font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
        <br>
      </div>
      <div dir="ltr"
style="font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
        Best — Juergen</div>
      <div dir="ltr"
style="font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
        <br>
      </div>
      <div dir="ltr"
style="font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
        <br>
      </div>
      <div dir="ltr"
style="font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
        <br>
      </div>
      <div id="ms-outlook-mobile-signature">
        <p
style="margin: 0in; font-family: Calibri, sans-serif; font-size: 11pt;"><span
style="font-family: Helvetica; font-size: 9pt; 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-family: Helvetica; font-size: 9pt; color: rgb(0, 120, 212);"><u><a
                href="mailto:jb77@buffalo.edu"
                title="mailto:jb77@buffalo.edu"
                data-outlook-id="4905ccdf-3cd0-4f5b-b56c-c0fede9584e6"
style="color: rgb(0, 120, 212); margin-top: 0px; margin-bottom: 0px;"
                moz-do-not-send="true" class="moz-txt-link-freetext">jb77@buffalo.edu</a></u></span><span
style="font-family: Helvetica; font-size: 9pt; color: black;"><br>
            Web: </span><span
style="font-family: Helvetica; font-size: 9pt; color: rgb(5, 99, 193);"><u><a
                href="http://www.acsu.buffalo.edu/~jb77/"
                title="http://www.acsu.buffalo.edu/~jb77/"
                data-outlook-id="3e486a1e-803e-4429-8f6b-ef75d02790e2"
style="color: rgb(5, 99, 193); margin-top: 0px; margin-bottom: 0px;"
                moz-do-not-send="true" class="moz-txt-link-freetext">http://www.acsu.buffalo.edu/~jb77/</a></u></span><span
style="font-family: Helvetica; font-size: 9pt; color: black;"> <br>
            <br>
          </span><span style="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-family: Helvetica; font-size: 9pt; color: black;"><br>
            <br>
            There’s A Crack In Everything - That’s How The Light Gets
            In <br>
            (Leonard Cohen)  </span></p>
        <p
style="margin: 0in; font-family: Calibri, sans-serif; font-size: 11pt;">-- </p>
        <p
style="margin: 0in; font-family: Calibri, sans-serif; font-size: 11pt;"> </p>
      </div>
      <div id="mail-editor-reference-message-container">
        <div class="ms-outlook-mobile-reference-message skipProofing"
style="text-align: left; padding: 3pt 0in 0in; border-width: 1pt medium medium; border-style: solid none none; border-color: rgb(181, 196, 223) currentcolor currentcolor; font-family: Aptos; font-size: 12pt; color: black;">
          <b>From: </b>Lingtyp
          <a class="moz-txt-link-rfc2396E" href="mailto:lingtyp-bounces@listserv.linguistlist.org"><lingtyp-bounces@listserv.linguistlist.org></a> on behalf of
          Peter Arkadiev via Lingtyp
          <a class="moz-txt-link-rfc2396E" href="mailto:lingtyp@listserv.linguistlist.org"><lingtyp@listserv.linguistlist.org></a><br>
          <b>Date: </b>Friday, November 21, 2025 at 05:59<br>
          <b>To: </b>Martin Haspelmath
          <a class="moz-txt-link-rfc2396E" href="mailto:martin_haspelmath@eva.mpg.de"><martin_haspelmath@eva.mpg.de></a>, Linguistic Typology
          <a class="moz-txt-link-rfc2396E" href="mailto:lingtyp@listserv.linguistlist.org"><lingtyp@listserv.linguistlist.org></a><br>
          <b>Subject: </b>Re: [Lingtyp] Reporting cross-linguistic
          frequencies<br>
          <br>
        </div>
        <div class="ms-outlook-mobile-reference-message skipProofing">Dear
          Martin, dear all,</div>
        <div class="ms-outlook-mobile-reference-message skipProofing"> </div>
        <div class="ms-outlook-mobile-reference-message skipProofing">I
          am starting to wonder whether statistical analysis of a
          language sample is at all a suitable method for "detecting
          universal tendencies that are caused by
          universal/non-historical factors" (Martin's formulation).
          Given that there is no consensus as for how to overcome
          genealogical and areal biases and even whether those biases
          must be overcome at all and what trying to overcome them
          actually gets us (apart from getting some of us high-profile
          publications with ever more complicated mathematical apparatus
          which others among us struggle to understand and cannot
          evaluate; not being in any way a "mathematically-gifted
          person", to borrow Stela's expression, I belong to the latter
          group), the whole enterprise does not appear to be very
          productive. What if the more appropriate method, at least if
          purported functional factors are being concerned, is the one
          employed by John Hawkins, Natalia Levshina and some others,
          i.e. to combine experimental research on production /
          processing with a quantitative study of variation in corpora
          across a small number of sufficiently distinct languages? If
          we can show that certain well-defined factors are operative in
          language processing and result in skewed distributions in
          corpora ultimately translatable into tendencies of diachronic
          change, and moreover are able to corroborate these results by
          similarly skewed distributions of variables in reasonably
          designed cross-linguistic samples, then what else do we need?
          In any case, as has been repeatedly stated many times, even if
          we find that in a certain language sample, however
          well-designed, a certain variable shows a clearly skewed
          distribution of, say 80% vs 20%, nothing follows from this in
          terms of "universal preferences" unless we are able to
          independently show that the more frequent value is in some or
          other way "preferred" in processing / production etc. I am
          sorry if the above is self-evident or naive.</div>
        <div class="ms-outlook-mobile-reference-message skipProofing"> </div>
        <div class="ms-outlook-mobile-reference-message skipProofing">Best
          regards,</div>
        <div class="ms-outlook-mobile-reference-message skipProofing"> </div>
        <div class="ms-outlook-mobile-reference-message skipProofing">Peter<br>
           </div>
        <div class="ms-outlook-mobile-reference-message skipProofing"> </div>
        <div class="ms-outlook-mobile-reference-message skipProofing">----------------</div>
        <div class="ms-outlook-mobile-reference-message skipProofing">Кому:
          <a class="moz-txt-link-abbreviated" href="mailto:lingtyp@listserv.linguistlist.org">lingtyp@listserv.linguistlist.org</a>
          (<a class="moz-txt-link-abbreviated" href="mailto:lingtyp@listserv.linguistlist.org">lingtyp@listserv.linguistlist.org</a>);</div>
        <div class="ms-outlook-mobile-reference-message skipProofing">Тема:
          [Lingtyp] Reporting cross-linguistic frequencies;</div>
        <div class="ms-outlook-mobile-reference-message skipProofing">21.11.2025,
          10:19, "Martin Haspelmath via Lingtyp"
          <a class="moz-txt-link-rfc2396E" href="mailto:lingtyp@listserv.linguistlist.org"><lingtyp@listserv.linguistlist.org></a>:</div>
        <blockquote>
          <p class="ms-outlook-mobile-reference-message skipProofing">Thanks,
            Jürgen! I like the "wave vs. particle" analogy, because
            these concrete expressions help us make sense of what seems
            to be going on (given the experimental results).</p>
          <p class="ms-outlook-mobile-reference-message skipProofing">In
            worldwide comparative linguistics, we also want to make
            sense of what is going on, but it seems to me that we need
            analogies not only for interpreting results, but also for
            understanding what we are aiming for. For me, "removing
            areal and genealogical/phylogenetic bias" has the aim of
            detecting universal tendencies that are caused by
            universal/non-historical factors.</p>
          <p class="ms-outlook-mobile-reference-message skipProofing">I
            would think that on the imagined concrete scenario of a
            sample of isolated isolates (e.g. 100 languages that have
            long existed on isolated islands, maybe of the Rapanui
            type), looking at these 100 isolates should give the same
            results as looking at 100 sample languages from larger
            families that have been shaped also by contact.</p>
          <p class="ms-outlook-mobile-reference-message skipProofing">Are
            there reasons to doubt this? If not, then we can take the
            "isolated isolates" scenario simply as a way of illustrating
            our goals in concrete terms (somewhat like "wave" and
            "particle" serve as concrete illustrations). </p>
          <p class="ms-outlook-mobile-reference-message skipProofing">But
            maybe the imagined scenario (which is not an "assumption"!!)
            is somehow problematic, because the goals of our enterprise
            are DIFFERENT. In Bickel's (2007) paper (LiTy 11), which has
            been widely cited, the idea seems to be that looking for
            "history-free" tendencies is somehow an obsolete goal.</p>
          <p class="ms-outlook-mobile-reference-message skipProofing">Some
            people have suggested that in identifying universal trends,
            one MUST take into account genealogies, and isolates are
            problematic because they are not part of any genealogy. This
            is because we should not look primarily at languages, but at
            *transitions* (changes from one type to another). If I
            understood Verkerk et al. (2025) correctly, they solved the
            "isolates problem" by using an artificial world tree (where
            all languages are somehow included; the very beautiful tree
            is used in <a
href="https://www.mpg.de/25723124/1114-evan-enduring-patterns-in-the-world-s-languages-150495-x"
              rel="noopener noreferrer"
originalsrc="https://www.mpg.de/25723124/1114-evan-enduring-patterns-in-the-world-s-languages-150495-x"
              data-outlook-id="12559ee5-fc85-478f-9674-ab4dfb981519"
              style="margin-top: 0px; margin-bottom: 0px;"
              moz-do-not-send="true">
              the press release</a>). Are Verkerk et al. pursuing a
            different goal? That is not really clear to me.</p>
          <p class="ms-outlook-mobile-reference-message skipProofing">I
            find the notion of an artificial world tree profoundly
            strange, much stranger than the hypothetical scenario of 100
            isolates on remote islands. But maybe it is needed, because
            the goal of the enterprise is somehow different (along
            Bickel's lines)? So I like the imagined "isolated isolates"
            scenario also because it clarifies what I'm interested in.</p>
          <p class="ms-outlook-mobile-reference-message skipProofing">(And
            isn't Trudgill's idea that isolates are somehow "exotic"
            very speculative? Shcherbakova et al. 2023 have not provided
            strong evidence against the idea, but they simply did not
            find evidence in favour of it.)</p>
          <p class="ms-outlook-mobile-reference-message skipProofing">One
            last point: Yes, all isolates are survivors from some larger
            family, but why is that relevant? Languages may have existed
            for half a million years or longer, and we know almost
            nothing about that deep past. Most of the currently existing
            families probably had more branches in earlier times, and
            the protolanguages we reconstruct may or may not have been
            isolates themselves. We cannot tell, but I don't see why we
            would need to know.</p>
          <p class="ms-outlook-mobile-reference-message skipProofing">Best,</p>
          <p class="ms-outlook-mobile-reference-message skipProofing">Martin</p>
          <p class="ms-outlook-mobile-reference-message skipProofing"> </p>
          <div class="ms-outlook-mobile-reference-message skipProofing">On
            21.11.25 07:07, Juergen Bohnemeyer via Lingtyp wrote:</div>
          <blockquote>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
              Dear all — Here’s a quick explanation of why the
              assumption of an “isolated isolate” is profoundly
              strange: </div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
              <br>
            </div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; color: rgb(0, 0, 0);">
              <span style="font-size: 12pt;">Leaving aside sign
                languages, constructed languages, and artificial
                languages, nobody seems to entertain the possibility
                that languages have emerged spontaneously out of
                something that we wouldn’t consider a language itself
                over the last few thousands of years. In other words,
                the languages we consider isolates are without exception
                lone survivors; but they did descend from  ancestors
                which are often
              </span><span
style="font-size: 16px; background-color: rgb(255, 255, 255);">lost and
                unknown</span><span style="font-size: 12pt;">, and these
                ancestors biased the offshoot's properties by dint of
                inheritance/transmission.</span></div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
              <br>
            </div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
              The reason isolates are interesting from a sampling
              perspective is that they may represent entire genera or
              families without forcing us to pick a member. But being an
              isolate does not mean being free of phylogenetic bias. On
              the contrary: isolates of unknown descend are actually
              particularly problematic in the sense that they are shaped
              by biases that we have no way of identifying directly
              since the biasing ancestors have been lost to time.</div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
              <br>
            </div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; color: rgb(0, 0, 0);">
              <span style="font-size: 12pt;">As to contact. Languages
                that are not in contact with other languages over long
                stretches of time are extremely rare and unusual. In
                fact, as I’m sure everyone here is aware, such languages
                have been plausibly argued to tend to evolve exotic
                properties as a result of their isolation (</span><span
style="font-size: 16px; background-color: rgb(255, 255, 255);">Lupyan
                & Dale 2010;
              </span><span style="font-size: 12pt;">Trudgill 2011),
                although this is controversial (Shcherbakova et al.
                2023). In any case, I would certainly not want to make
                such languages the basis for causal inference in
                typology.</span></div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
              <br>
            </div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
              But it gets a lot worse. The “isolated isolate”
              interpretation doesn’t just require us to think of a
              language that isn’t currently in contact with any other
              language. We would have to assume a language that has
              <b>never</b>​ come into contact with any other language at
              any point in its history (at least not long/intensively
              enough to change as a result of it). I’m seriously
              uncertain whether such a language has ever existed on this
              planet. </div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
              <br>
            </div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
              Here’s an analogy from quantum mechanics: Schrödinger’s
              and Heisenberg’s equations are mathematical models that
              describe the experimentally observed behavior of
              elementary particles under various conditions. The
              particle and the wave interpretation are interpretations
              that we use to make sense of these mathematical models. We
              find these models useful because most of us don’t think in
              mathematical equations (not even theoretical physicists,
              it would seem). But if we push these interpretations
              beyond a certain point, they break down. To begin with, we
              can’t think of something simultaneously as a wave and as a
              particle. </div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
              <br>
            </div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
              In the same way, we can mathematically describe the
              influence phylogeny and areality exert on the probability
              of a particular language having certain properties. The
              “isolated isolate” interpretation is just that - an
              interpretation of the statistical models; but, as I tried
              to show above, it runs into absurdities rather more
              quickly than the particle and wave interpretations in
              quantum mechanics. </div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
              <br>
            </div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
              Best — Juergen</div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
              <br>
            </div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
              G. Lupyan, R. Dale, Language structure is partly
              determined by social structure. PLOS ONE5, e8559 (2010).</div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
              <br>
            </div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
              O. Shcherbakova, S. M. Michaelis, H. J. Haynie, et al.
              Societies of strangers do not speak less complex
              languages.
              <i>Scientific Advances </i>9, eadf7704 (2023).</div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
              <br>
            </div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
              P. Trudgill, <i>Sociolinguistic Typology: Social
                Determinants of Linguistic Complexity
              </i>(OxfordUniv. Press, 2011).</div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"
style="font-family: aptos, arial, helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
              <br>
            </div>
            <div
id="1dadbb124ddfd57c2650379a47ad6a82ms-outlook-mobile-signature">
              <p
style="margin: 0in; font-family: calibri, sans-serif; font-size: 11pt;"><span
style="font-family: helvetica; font-size: 9pt; 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-family: helvetica; font-size: 9pt; color: rgb(0, 120, 212);"><u><a
                      href="mailto:jb77@buffalo.edu"
class="56221ecd4cd88a7e220fd42e552d23b7moz-txt-link-freetext moz-txt-link-freetext"
                      title="mailto:jb77@buffalo.edu"
data-outlook-id="258169d3-1e7c-4d8b-966b-74a64ce0e7f9"
style="color: rgb(0, 120, 212); margin-top: 0px; margin-bottom: 0px;"
                      moz-do-not-send="true">jb77@buffalo.edu</a></u></span><span
style="font-family: helvetica; font-size: 9pt; color: black;"><br>
                  Web: </span><span
style="font-family: helvetica; font-size: 9pt; color: rgb(5, 99, 193);"><u><a
                      href="http://www.acsu.buffalo.edu/~jb77/"
class="56221ecd4cd88a7e220fd42e552d23b7moz-txt-link-freetext moz-txt-link-freetext"
                      title="http://www.acsu.buffalo.edu/~jb77/"
data-outlook-id="7788a104-a393-4a98-9a08-c29ad6c6cab0"
style="color: rgb(5, 99, 193); margin-top: 0px; margin-bottom: 0px;"
                      moz-do-not-send="true">http://www.acsu.buffalo.edu/~jb77/</a></u></span><span
style="font-family: helvetica; font-size: 9pt; color: black;"> <br>
                  <br>
                </span><span style="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-family: helvetica; font-size: 9pt; color: black;"><br>
                  <br>
                  There’s A Crack In Everything - That’s How The Light
                  Gets In <br>
                  (Leonard Cohen)  </span></p>
              <p
style="margin: 0in; font-family: calibri, sans-serif; font-size: 11pt;">-- </p>
              <p
style="margin: 0in; font-family: calibri, sans-serif; font-size: 11pt;"> </p>
            </div>
            <div
id="a072bc9780e9b7b3b9a49a25572a5efdmail-editor-reference-message-container">
              <div
class="39e1103e8e8cbb86b1168768346b5522ms-outlook-mobile-reference-message 728e41e5405915cc550e9daf9431b8d6skipProofing"
style="text-align: left; padding: 3pt 0in 0in; border-width: 1pt medium medium; border-style: solid none none; border-color: rgb(181, 196, 223) currentcolor currentcolor; font-family: aptos; font-size: 12pt; color: black;">
                <b>From: </b>Lingtyp <a
href="mailto:lingtyp-bounces@listserv.linguistlist.org"
class="1f85fe41a9161661477b40489dd2f552moz-txt-link-rfc2396E"
                  data-outlook-id="e04ccbed-155f-4581-9d66-c5563eba94e4"
                  moz-do-not-send="true">
                  <lingtyp-bounces@listserv.linguistlist.org></a> on
                behalf of Matías Guzmán Naranjo via Lingtyp
                <a href="mailto:lingtyp@listserv.linguistlist.org"
class="1f85fe41a9161661477b40489dd2f552moz-txt-link-rfc2396E"
                  data-outlook-id="6229b204-2ed7-4bef-9e30-29feba93f6ab"
                  moz-do-not-send="true">
                  <lingtyp@listserv.linguistlist.org></a><br>
                <b>Date: </b>Thursday, November 20, 2025 at 04:01<br>
                <b>To: </b><a
                  href="mailto:lingtyp@listserv.linguistlist.org"
class="324de92b3f6b2f5e993df2fdf11fa1c7moz-txt-link-abbreviated moz-txt-link-freetext"
                  data-outlook-id="09061d23-fa38-4dce-9b0f-a433159fb284"
                  moz-do-not-send="true">lingtyp@listserv.linguistlist.org</a>
                <a href="mailto:lingtyp@listserv.linguistlist.org"
class="1f85fe41a9161661477b40489dd2f552moz-txt-link-rfc2396E"
                  data-outlook-id="18e364c6-0953-4759-a7ff-229a76b2e83c"
                  moz-do-not-send="true">
                  <lingtyp@listserv.linguistlist.org></a><br>
                <b>Subject: </b>Re: [Lingtyp] Reporting
                cross-linguistic frequencies<br>
                <br>
              </div>
              <div class="662abf391f8b56b11ba8165b7bd48bb4PlainText"
                style="font-size: 11pt;">I'll jump in with some
                thoughts.<br>
                <br>
                <br>
                - Dryer's method and ours aim at doing basically the
                same thing:<br>
                quantifying what's "left" after removing genetic and
                areal bias.<br>
                <br>
                - Whether you should call them proportions or adjusted
                frequencies...<br>
                I'm not sure that it matters that much? As long as you
                understand how<br>
                they were calculated...<br>
                <br>
                - How you want to interpret this "what's left" is
                debatable, I guess,<br>
                but I don't think I agree with Jürgen. As far as I can
                tell it should be<br>
                compatible with something along the lines of an
                "isolated isolate" as<br>
                described by Martin. You can also see them as
                'universal' preferences<br>
                (more or less the same thing?).<br>
                <br>
                - "the probability of a random language having a certain
                property<br>
                depends on (or is influenced by, or varies with, etc.)
                it being related<br>
                to certain other languages, or being  spoken (or signed)
                in a particular<br>
                area" -> In our approach we assumes that the
                probability of a language L<br>
                having some feature value F depends on three things: 1)
                its relatedness<br>
                to other languages, 2) its contact to other languages,
                3) some universal<br>
                preference for F. Kind of the point of what we do is
                that we try to<br>
                estimate each of these factors. [We can add more factors
                and more<br>
                structure, but that's the most basic model]<br>
                <br>
                - You can quantify the contribution of the phylogenetic
                component and<br>
                the areal component(s) with our techniques, but this is
                a bit tricky<br>
                because there is unavoidable overlap in the information
                each one<br>
                contains. These measures also have a different meaning
                than the adjusted<br>
                frequency and can't be used as a replacement for them,
                you can use them<br>
                in addition to.<br>
                <br>
                <br>
                Matías<br>
                <br>
                <br>
                <br>
                El 20/11/25 a las 9:36, Omri Amiraz via Lingtyp
                escribió:<br>
                > Dear all,<br>
                > I agree with Ian that, in addition to genealogical
                and areal biases,<br>
                > the very question of what counts as a language
                versus a dialect is<br>
                > partly subjective. This makes actual frequencies
                even more<br>
                > problematic, since we would obtain different
                results depending on<br>
                > whether we treat Wu Chinese as one language or as
                thirty separate<br>
                > languages, as Ian pointed out.<br>
                > Juergen wrote: "We can empirically assess the
                extent to which the<br>
                > probability of a random language having a certain
                property depends on<br>
                > (or is influenced by, or varies with, etc.) it
                being related to<br>
                > certain other languages, or being  spoken (or
                signed) in a particular<br>
                > area."<br>
                ><br>
                > I wonder whether it might be useful to have a
                measure of the<br>
                > genealogical and areal spread of a feature,
                essentially quantifying<br>
                > how broadly it is distributed across families and
                regions in the<br>
                > present-day world. Such a measure might be more
                straightforward to<br>
                > interpret than an adjusted frequency/probability,
                since it is not<br>
                > clear whether the described population is a
                hypothetical set of<br>
                > isolated isolates or something else.<br>
                ><br>
                > Is anyone aware of an existing metric that captures
                genealogical or<br>
                > areal spread in this way?<br>
                ><br>
                > Best,<br>
                > Omri<br>
                ><br>
                > _______________________________________________<br>
                > Lingtyp mailing list<br>
                > <a href="mailto:Lingtyp@listserv.linguistlist.org"
class="324de92b3f6b2f5e993df2fdf11fa1c7moz-txt-link-abbreviated moz-txt-link-freetext"
                  data-outlook-id="b0d26597-ae71-45b6-86fa-086108b1854a"
                  moz-do-not-send="true">
                  Lingtyp@listserv.linguistlist.org</a><br>
                > <a
href="https://listserv.linguistlist.org/cgi-bin/mailman/listinfo/lingtyp"
originalsrc="https://listserv.linguistlist.org/cgi-bin/mailman/listinfo/lingtyp"
                  data-outlook-id="a5a3d817-9c03-4c2e-8a5b-44f65180460a"
                  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%7C02%7Cjb77%40buffalo.edu%7C88b1df86321b4cb12f9f08de28135c96%7C96464a8af8ed40b199e25f6b50a20250%7C0%7C0%7C638992260962407959%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=uY52%2BPtTVyzNB0LIowvZ0UzKWB6MWLR%2BG62V70JtNGE%3D&reserved=0</a><br>
                _______________________________________________<br>
                Lingtyp mailing list<br>
                <a href="mailto:Lingtyp@listserv.linguistlist.org"
class="324de92b3f6b2f5e993df2fdf11fa1c7moz-txt-link-abbreviated moz-txt-link-freetext"
                  data-outlook-id="b5c40087-9974-4b0d-91de-ef125dafb5af"
                  moz-do-not-send="true">Lingtyp@listserv.linguistlist.org</a><br>
                <a
href="https://listserv.linguistlist.org/cgi-bin/mailman/listinfo/lingtyp"
originalsrc="https://listserv.linguistlist.org/cgi-bin/mailman/listinfo/lingtyp"
                  data-outlook-id="09dcf689-cd6a-410c-ae24-2a4d6a68593a"
                  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%7C02%7Cjb77%40buffalo.edu%7C88b1df86321b4cb12f9f08de28135c96%7C96464a8af8ed40b199e25f6b50a20250%7C0%7C0%7C638992260962443120%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=X%2F1JMgRNS%2Bn0ZlGa7pPdsJWJBoJy%2BYJt6bHWktCMeRc%3D&reserved=0</a></div>
            </div>
            <div dir="ltr"
              class="ms-outlook-mobile-reference-message skipProofing"><br>
            </div>
            <pre><div
            class="3f7f1cfb43cdc145acb8dd7a82f3a2c8moz-quote-pre">_______________________________________________
Lingtyp mailing list
<a href="mailto:Lingtyp@listserv.linguistlist.org"
class="324de92b3f6b2f5e993df2fdf11fa1c7moz-txt-link-abbreviated moz-txt-link-freetext"
            data-outlook-id="a0be30c7-f0d3-43dd-abbc-bd1e28d647e8"
            moz-do-not-send="true">Lingtyp@listserv.linguistlist.org</a>
<a
href="https://listserv.linguistlist.org/cgi-bin/mailman/listinfo/lingtyp"
class="56221ecd4cd88a7e220fd42e552d23b7moz-txt-link-freetext moz-txt-link-freetext"
originalsrc="https://listserv.linguistlist.org/cgi-bin/mailman/listinfo/lingtyp"
            data-outlook-id="ccabecf4-c84b-4248-aefd-e19402b972a0"
            moz-do-not-send="true">https://listserv.linguistlist.org/cgi-bin/mailman/listinfo/lingtyp</a>
</div></pre>
          </blockquote>
          <pre><div
          class="ms-outlook-mobile-reference-message skipProofing">-- 
Martin Haspelmath
Max Planck Institute for Evolutionary Anthropology
Deutscher Platz 6
D-04103 Leipzig
<a
href="https://www.eva.mpg.de/linguistic-and-cultural-evolution/staff/martin-haspelmath/"
          rel="noopener noreferrer"
originalsrc="https://www.eva.mpg.de/linguistic-and-cultural-evolution/staff/martin-haspelmath/"
          data-outlook-id="6ea7f369-3a66-434a-97f2-1121bcf31f88"
          moz-do-not-send="true" class="moz-txt-link-freetext">https://www.eva.mpg.de/linguistic-and-cultural-evolution/staff/martin-haspelmath/</a></div></pre>
          <div class="ms-outlook-mobile-reference-message skipProofing">,</div>
          <p class="ms-outlook-mobile-reference-message skipProofing">_______________________________________________<br>
            Lingtyp mailing list<br>
            <a href="mailto:Lingtyp@listserv.linguistlist.org"
              rel="noopener noreferrer"
              data-outlook-id="57253f9a-b88d-498c-8587-8933cf07fc07"
              style="margin-top: 0px; margin-bottom: 0px;"
              moz-do-not-send="true" class="moz-txt-link-freetext">Lingtyp@listserv.linguistlist.org</a><br>
            <a
href="https://listserv.linguistlist.org/cgi-bin/mailman/listinfo/lingtyp"
              rel="noopener noreferrer"
originalsrc="https://listserv.linguistlist.org/cgi-bin/mailman/listinfo/lingtyp"
              data-outlook-id="7655b7dc-2ade-4059-8456-51e83c5962d7"
              style="margin-top: 0px; margin-bottom: 0px;"
              moz-do-not-send="true" class="moz-txt-link-freetext">https://listserv.linguistlist.org/cgi-bin/mailman/listinfo/lingtyp</a></p>
        </blockquote>
        <div class="ms-outlook-mobile-reference-message skipProofing"> </div>
        <div class="ms-outlook-mobile-reference-message skipProofing"> </div>
        <div class="ms-outlook-mobile-reference-message skipProofing">-- </div>
        <div class="ms-outlook-mobile-reference-message skipProofing">Peter
          Arkadiev, PhD Habil.</div>
        <div class="ms-outlook-mobile-reference-message skipProofing"><a class="moz-txt-link-freetext" href="https://peterarkadiev.github.io/">https://peterarkadiev.github.io/</a></div>
        <div class="ms-outlook-mobile-reference-message skipProofing"> </div>
      </div>
    </blockquote>
    <pre class="moz-signature" cols="72">-- 
Martin Haspelmath
Max Planck Institute for Evolutionary Anthropology
Deutscher Platz 6
D-04103 Leipzig
<a class="moz-txt-link-freetext" href="https://www.eva.mpg.de/linguistic-and-cultural-evolution/staff/martin-haspelmath/">https://www.eva.mpg.de/linguistic-and-cultural-evolution/staff/martin-haspelmath/</a></pre>
  </body>
</html>