<html>
  <head>
    <meta content="text/html; charset=windows-1252"
      http-equiv="Content-Type">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    Thanks, Östen, for this further discussion of boundness, and for
    digging into Bloomfield!<br>
    <br>
    I'm not saying that Bloomfield was right, but he was the first to
    propose a serious definition of "word" – so even though he failed,
    it's still instructive to read him.<br>
    <br>
    In any event, his "bound vs. free" distinction survived, and I think
    we need it for the definition of some key concepts in typology:<br>
    <br>
    – auxiliary verb vs. tense-aspect marker<br>
    – adposition vs. relational noun<br>
    – person index vs. free person pronoun <br>
    <br>
    Linguists often try to define grammatical elements in terms of their
    "grammatical" meanings, but it seems to me that a grammatical marker
    in general is best defined as "an element that cannot occur in
    isolation".<br>
    <br>
    (Of course, this is paradigmatic in the sense that it relies on
    "trans-utterance equivalence", but this applies to every concept in
    grammar, as far as I can see.)<br>
    <br>
    The nice thing about Bloomfield's "bound vs. free" distinction is
    that it's a single criterion (rather than a battery of criteria, so
    that the issue of disjunctive definitions doesn't apply), and that
    it's universally applocable, because all languages have free forms
    and bound forms (of course with some questionable cases in the
    middle, as is always the case).<br>
    <br>
    The term "bound" is also often used in a phonological sense
    ("phonologically bound"), but there are many different ways in which
    this can be interpreted, so I find the Bloomfieldian sense much more
    readily applicable.<br>
    <br>
    (Strangely, I often read linguists talk about "bound morphology", or
    "bound affixes" – I wonder whether this means anything, or whether
    it's just sloppy usage.)<br>
    <br>
    Best,<br>
    Martin<br>
    <br>
    <div class="moz-cite-prefix">On 18.11.17 18:31, Östen Dahl wrote:<br>
    </div>
    <blockquote
      cite="mid:1849537f12604652b7f736a20a70823a@ebox-prod-srv10.win.su.se"
      type="cite">
      <meta http-equiv="Content-Type" content="text/html;
        charset=windows-1252">
      <meta name="Generator" content="Microsoft Word 15 (filtered
        medium)">
      <style><!--
/* Font Definitions */
@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:Consolas;
        panose-1:2 11 6 9 2 2 4 3 2 4;}
@font-face
        {font-family:Times;
        panose-1:2 2 6 3 5 4 5 2 3 4;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
        {margin:0cm;
        margin-bottom:.0001pt;
        font-size:11.0pt;
        font-family:"Calibri",sans-serif;}
a:link, span.MsoHyperlink
        {mso-style-priority:99;
        color:blue;
        text-decoration:underline;}
a:visited, span.MsoHyperlinkFollowed
        {mso-style-priority:99;
        color:purple;
        text-decoration:underline;}
pre
        {mso-style-priority:99;
        mso-style-link:"HTML - förformaterad Char";
        margin:0cm;
        margin-bottom:.0001pt;
        font-size:10.0pt;
        font-family:"Courier New";}
p.msonormal0, li.msonormal0, div.msonormal0
        {mso-style-name:msonormal;
        mso-margin-top-alt:auto;
        margin-right:0cm;
        mso-margin-bottom-alt:auto;
        margin-left:0cm;
        font-size:11.0pt;
        font-family:"Calibri",sans-serif;}
span.HTML-frformateradChar
        {mso-style-name:"HTML - förformaterad Char";
        mso-style-priority:99;
        mso-style-link:"HTML - förformaterad";
        font-family:Consolas;}
span.hoenzb
        {mso-style-name:hoenzb;}
span.E-postmall22
        {mso-style-type:personal;
        font-family:"Calibri",sans-serif;
        color:windowtext;}
span.E-postmall23
        {mso-style-type:personal-compose;
        font-family:"Calibri",sans-serif;
        color:windowtext;}
.MsoChpDefault
        {mso-style-type:export-only;
        font-size:10.0pt;}
@page WordSection1
        {size:612.0pt 792.0pt;
        margin:70.85pt 70.85pt 70.85pt 70.85pt;}
div.WordSection1
        {page:WordSection1;}
--></style><!--[if gte mso 9]><xml>
<o:shapedefaults v:ext="edit" spidmax="1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext="edit">
<o:idmap v:ext="edit" data="1" />
</o:shapelayout></xml><![endif]-->
      <div class="WordSection1">
        <p class="MsoNormal"><span
            style="font-size:12.0pt;mso-fareast-language:EN-US"
            lang="EN-US">I want to comment on the notion of "boundness",
            in particular in its Bloomfieldian version. Matthew has
            already expressed some doubts about the applicability of
            this notion, and I was thinking of some further problems --
            then I went to the source (Bloomfield's Language from 1933)
            and found that (perhaps not unexpectedly) Bloomfield had
            already considered those problems. In fact, in addition to
            the definition on p. 160, he devotes at least seven pages to
            discussing them (177-184). But the way he does so does not
            really solve the problems with his definitions, they rather
            confound them even more.<o:p></o:p></span></p>
        <p class="MsoNormal"><span
            style="font-size:12.0pt;mso-fareast-language:EN-US"
            lang="EN-US"><o:p> </o:p></span></p>
        <p class="MsoNormal"><span
            style="font-size:12.0pt;mso-fareast-language:EN-US"
            lang="EN-US">The definitions of "bound" and "free" given on
            p. 160 are clear enough: "A linguistic form which is never
            spoken alone is a *bound* form; all others (as, for
            instance, *John ran* or *John* or *run* or *running*) are
            *free* forms." Furthermore, we learn on p. 178 that a word
            is a "minimum free form", i.e. a free form which does not
            consist entirely of two or more lesser free forms.
            <o:p></o:p></span></p>
        <p class="MsoNormal"><span
            style="font-size:12.0pt;mso-fareast-language:EN-US"
            lang="EN-US"><o:p> </o:p></span></p>
        <p class="MsoNormal"><span
            style="font-size:12.0pt;mso-fareast-language:EN-US"
            lang="EN-US">It is not quite clear, however, what Bloomfield
            really wants to say about boundness and wordhood. On p. 179,
            he says that "[i]n the case of many languages, it is
            impossible to distinguish consistently, on the one hand,
            between phrases and words and, on the other hand, between
            words and bound forms." On p. 181, he likewise notes that
            "many forms lie on the border-line between bound forms and
            words, or between words and phrases; it is impossible to
            make a rigid distinction between forms that may and forms
            that may not be spoken in absolute position". Depending on
            how rigid we want our distinctions to be, this may seem to
            be a more or less serious objection to the notions of
            boundness and wordhood, but I find the cases more
            problematic where Bloomfield seems to want to diss his own
            definitions. Sometimes, he says, "the general structure of a
            language may make one classification more convenient for our
            purpose" (179). Thus the English definite article *the*
            fills the same syntactic slot as the demonstratives *this*
            and *that*. Therefore, even if *the* only occurs alone in
            "far-fetched situations", the parallelism with the
            demonstratives, which "freely occur as sentences", "leads us
            to class *the* as a word". Another example mentioned by
            Bloomfield concerns French pronouns such as *je* and *il*
            which usually do not appear in "absolute use", but which
            again, due to the parallelism with strong forms such as
            *moi* and *lui*, "have the status of words" (180).
            (Bloomfield does not say if this means that they also have
            the status of free forms, but given that wordhood was
            previously defined in terms of boundness, this is a
            plausible inference.)<o:p></o:p></span></p>
        <p class="MsoNormal"><span
            style="font-size:12.0pt;mso-fareast-language:EN-US"
            lang="EN-US"><o:p> </o:p></span></p>
        <p class="MsoNormal"><span
            style="font-size:12.0pt;mso-fareast-language:EN-US"
            lang="EN-US">What this shows is that Bloomfield had a rather
            pragmatic attitude to his notions and it is questionable if
            we should speak of "Bloomfieldian" concepts if it means that
            we want to be more Bloomfieldian than Bloomfield himself.
            But a question of greater theoretical significance is to
            what extent considerations of what Bloomfield refers to as
            the general structure of the language, or of questions
            regarding the optimal description of an individual language
            system, is allowed to influence the application of
            comparative concepts. Martin's work on those concepts seems
            to imply that there should be no such influence. He says
            (Haspelmath 2010:680f.): "Comparative concepts are motivated
            and defined in a way that is quite independent of linguistic
            categories (though of course not independent of the facts of
            languages)" and "[i]n practice, typologists do not
            generalize over the categories of languages, but over
            properties of languages that they identify regardless of the
            categories that speakers seem to have internalized and that
            structural analysis reveals". I think that what "structural
            analysis reveals" is precisely what Bloomfield was talking
            about.
            <o:p></o:p></span></p>
        <p class="MsoNormal"><span
            style="font-size:12.0pt;mso-fareast-language:EN-US"
            lang="EN-US"><o:p> </o:p></span></p>
        <p class="MsoNormal"><span
            style="font-size:12.0pt;mso-fareast-language:EN-US"
            lang="EN-US">In my opinion, what is crucial here is that the
            Bloomfieldian definitions of boundness and wordhood are
            paradigmatic in the sense that they demand the consideration
            of more than one utterance, and thus involve making
            judgments about trans-utterance and trans-construction
            identity and equivalence. Thus, we have to make up our mind
            about whether *je* and *moi* are the same entity or not. If
            we think the answer to that question is clear, there are
            many similar cases which are tougher to decide. And the way
            we usually make such decisions is to consider what is the
            best way to capture the general structure of the language,
            or what "structural analysis reveals". In other words,
            Bloomfieldian boundness is dependent on language-particular
            descriptions and categories.    <o:p>
            </o:p></span></p>
        <p class="MsoNormal"><span
            style="font-size:12.0pt;mso-fareast-language:EN-US"
            lang="EN-US"><o:p> </o:p></span></p>
        <p class="MsoNormal"><span
            style="font-size:12.0pt;mso-fareast-language:EN-US"
            lang="EN-US">Bloomfield, Leonard. 1933. Language,. New York:
            H. Holt and Company.<o:p></o:p></span></p>
        <p class="MsoNormal"><span
            style="font-size:12.0pt;mso-fareast-language:EN-US"
            lang="EN-US"><o:p> </o:p></span></p>
        <p class="MsoNormal"><span
            style="font-size:12.0pt;mso-fareast-language:EN-US"
            lang="EN-US">Haspelmath, Martin. 2010. Comparative concepts
            and descriptive categories in crosslinguistic studies.
          </span><span
            style="font-size:12.0pt;mso-fareast-language:EN-US">Language
            86(3). 663–687.<o:p></o:p></span></p>
        <p class="MsoNormal"><span style="mso-fareast-language:EN-US"><o:p> </o:p></span></p>
        <div>
          <div style="border:none;border-top:solid #E1E1E1
            1.0pt;padding:3.0pt 0cm 0cm 0cm">
            <p class="MsoNormal"><b>Från:</b> Lingtyp
              [<a class="moz-txt-link-freetext" href="mailto:lingtyp-bounces@listserv.linguistlist.org">mailto:lingtyp-bounces@listserv.linguistlist.org</a>]
              <b>För </b>William Croft<br>
              <b>Skickat:</b> den 18 november 2017 17:38<br>
              <b>Till:</b> Volker Gast <a class="moz-txt-link-rfc2396E" href="mailto:volker.gast@uni-jena.de"><volker.gast@uni-jena.de></a>;
              Linguistic Typology
              <a class="moz-txt-link-rfc2396E" href="mailto:lingtyp@listserv.linguistlist.org"><lingtyp@listserv.linguistlist.org></a><br>
              <b>Ämne:</b> Re: [Lingtyp] wordhood: bonded vs. bound<o:p></o:p></p>
          </div>
        </div>
        <p class="MsoNormal"><o:p> </o:p></p>
        <div>
          <p class="MsoNormal">Dear Volker,<o:p></o:p></p>
        </div>
        <div>
          <p class="MsoNormal"><o:p> </o:p></p>
        </div>
        <div>
          <p class="MsoNormal">  I think most typologists are aware that
            (i) defining categories for coding is very hard, especially
            across languages -- hence all the discussions about
            comparative concepts on Lingtyp (some of which have
            subsequently been published in some form in Linguistic
            Typology), of which this discussion of ‘word’ is only the
            latest; and (ii) that typologists must usually
            operationalize those criteria and make the
            operationalizations as explicit as possible. I think that
            (i) and (ii) are fairly common practice in typology, despite
            my previous comments about essentialism and methodological
            opportunism (cherry-picking of criteria).<o:p></o:p></p>
        </div>
        <div>
          <p class="MsoNormal"><o:p> </o:p></p>
        </div>
        <div>
          <p class="MsoNormal">   On the other hand, your point about
            mono-annotator annotation is well taken. Nevertheless, the
            operational factor is this one:<o:p></o:p></p>
        </div>
        <blockquote style="margin-top:5.0pt;margin-bottom:5.0pt">
          <div>
            <p class="MsoNormal"
              style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">And
              I'm not saying that mono-annotator projects are useless,
              sometimes you just don't have the manpower for
              multi-annotator projects <o:p></o:p></p>
          </div>
        </blockquote>
        <div>
          <p class="MsoNormal">  I have recently been working on
            computational projects that involve annotation, and even
            there, where there is a lot more large-scale funding than in
            typology, it is very expensive to hire and train annotators,
            and in the end there are maybe two annotators and a third
            person acting as adjudicator for a pilot annotation at most.
            (In fact, most of the effort in computational linguistics is
            towards training classifiers to do the annotation
            automatically on large corpora, and in my small experience
            those are often worse than mono-annotator annotations.)<o:p></o:p></p>
        </div>
        <div>
          <p class="MsoNormal"><o:p> </o:p></p>
        </div>
        <div>
          <p class="MsoNormal">    In typology, there is virtually no
            funding for any sort of multi-annotator annotation
            whatsoever. This is especially true for graduate students
            doing typological dissertations, but also for faculty doing
            typological research. I would guess that many typologists
            are aware that multi-annotator annotation is preferable, but
            impractical. But we don’t normally add a statement like “We
            are aware that engaging multiple annotators would improve
            the reliability of our coding and hence of the results of
            our crosslinguistic study; but due to lack of funding, all
            annotation of the data was performed by the author.” Perhaps
            we typologists should starting adding such statements.<o:p></o:p></p>
        </div>
        <div>
          <p class="MsoNormal"><o:p> </o:p></p>
        </div>
        <div>
          <p class="MsoNormal">Best wishes,<o:p></o:p></p>
        </div>
        <div>
          <p class="MsoNormal">Bill<o:p></o:p></p>
        </div>
        <div>
          <p class="MsoNormal"><br>
            <br>
            <o:p></o:p></p>
          <blockquote style="margin-top:5.0pt;margin-bottom:5.0pt">
            <div>
              <p class="MsoNormal">On Nov 18, 2017, at 6:32 AM, Volker
                Gast <<a moz-do-not-send="true"
                  href="mailto:volker.gast@uni-jena.de">volker.gast@uni-jena.de</a>>
                wrote:<o:p></o:p></p>
            </div>
            <p class="MsoNormal"><o:p> </o:p></p>
            <div>
              <div>
                <p class="MsoNormal"
                  style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><o:p> </o:p></p>
                <p class="MsoNormal"
                  style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">Hi
                  Johanna, even if I could do this diplomatically, I
                  wouldn't, and I think it wouldn't make much sense, as
                  my point is not about specific publications or
                  authors; it's about common practice (and common
                  practice is reflected in the publications of 'major
                  authorities'). But I think I get your point; so let me
                  be a bit more specific.<o:p></o:p></p>
                <p class="MsoNormal"
                  style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">A
                  lot of (quantitative) typological work relies on
                  'coding': Information is extracted from grammars and
                  transformed into a data matrix. Now, it is common
                  practice (and I'm not excluding myself here) for the
                  coding to be done by the analyst him/herself, and by
                  no one else. But that's considered bad practice in
                  other fields. Ideally, you'd need a team of annotators
                  coding independently, on the basis of annotation
                  guidelines. The team codes a sample, determines
                  inter-annotator agreement, and adjusts/specifies the
                  annotation guidelines where necessary. This is done
                  until the inter-annotator agreement is satisfactory.
                  And then you can start with the actual coding.
                  Ideally, the analyst shouldn't be involved in the
                  coding process, as her annotation decisions might be
                  (subconsciously) influenced by her working hypotheses.
                  (Note that this might be a viable solution to the
                  question of how comparative concepts can reliably be
                  defined, for a given study; you can just measure how
                  much inter-annotator variation there is; whether or
                  not the operationalizations make sense is a different
                  question, of course, one of validity. When you use a
                  set of criteria disjunctively, the question is what
                  exactly your operationalizations are intended to
                  represent.)<o:p></o:p></p>
                <p class="MsoNormal"
                  style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">Note
                  that I'm not saying that there are no multi-annotator
                  projects in typology (I'm actually involved in two
                  such projects, though one of them is actually a
                  comparative corpus linguistics project); but as far as
                  I can tell, it is 'basically' comon practice for
                  analysts to code the data themselves. And I'm not
                  saying that mono-annotator projects are useless,
                  sometimes you just don't have the manpower for
                  multi-annotator projects (and one of the
                  multi-annotator projects I'm involved in was really
                  painful; but it was instructive to see that even for
                  categories that we thought we had defined rather
                  clearly, inter-annotator agreement was rather low in
                  some cases). But as I said earlier, it would be nice
                  to have some standards or at least general guidelines
                  for coding typological data. Minimally, I think, the
                  data should be published, along with at least some
                  information on the operational tests that were
                  applied, even if done by a single annotator.<o:p></o:p></p>
                <p class="MsoNormal"
                  style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">I
                  hope this clarifies my (too general) remarks in my
                  previous post.<br>
                  Volker<o:p></o:p></p>
                <p class="MsoNormal"><o:p> </o:p></p>
                <div>
                  <p class="MsoNormal">Am 18.11.2017 um 13:27 schrieb
                    Johanna NICHOLS:<o:p></o:p></p>
                </div>
                <blockquote style="margin-top:5.0pt;margin-bottom:5.0pt">
                  <div>
                    <div>
                      <div>
                        <p class="MsoNormal"
                          style="margin-bottom:12.0pt">Volker,  <o:p></o:p></p>
                      </div>
                      <p class="MsoNormal">If there's a way to do this
                        diplomatically, could you cite an example or two
                        of  "important publications by major authorities
                        of the field where these criteria are simply not
                        applied"?   In linguistics we don't have as much
                        technical comment on publications as some other
                        fields do, and maybe we should.  In journals
                        where I see technical comments sections those
                        comments are refereed, edited, brief, and
                        focused on factual and methodological matters,
                        i.e. about empirical fundamentals and not debate
                        on theoretical frameworks.<o:p></o:p></p>
                    </div>
                    <div>
                      <p class="MsoNormal"><o:p> </o:p></p>
                    </div>
                    <div>
                      <p class="MsoNormal">If there's no way to do it
                        diplomatically, never mind.<o:p></o:p></p>
                    </div>
                    <div>
                      <p class="MsoNormal"><o:p> </o:p></p>
                    </div>
                    <p class="MsoNormal">Johanna<o:p></o:p></p>
                  </div>
                  <div>
                    <p class="MsoNormal"><o:p> </o:p></p>
                    <div>
                      <p class="MsoNormal">On Sat, Nov 18, 2017 at 12:37
                        PM, Volker Gast <<a moz-do-not-send="true"
                          href="mailto:volker.gast@uni-jena.de"
                          target="_blank">volker.gast@uni-jena.de</a>>
                        wrote:<o:p></o:p></p>
                      <blockquote style="border:none;border-left:solid
                        #CCCCCC 1.0pt;padding:0cm 0cm 0cm
                        6.0pt;margin-left:4.8pt;margin-right:0cm">
                        <div>
                          <p class="MsoNormal"
                            style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">Matthew
                            -- are you saying that "one cannot rule out
                            disjunctively defined comparative concept"
                            because this is what you did?<o:p></o:p></p>
                          <p class="MsoNormal"
                            style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">I
                            am not convinced by "disjunctive comparative
                            concepts". Now, that's nothing for you to
                            worry about -- I'm just one reader
                            (actually, audience of your ALT/2015-talk)
                            who doesn't buy your conclusions because he
                            doesn't accept your operationalizations.<o:p></o:p></p>
                          <p class="MsoNormal"
                            style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">But
                            if we want "to talk TO each other (not only
                            PAST each other)", as Martin writes, it
                            would be good to have what other fields call
                            "standards of empirical research". We have
                            copied a lot of statistical methods from
                            fields such as the social sciences and
                            biology. I think it would also be beneficial
                            to take a look at their standards at the
                            "lower" level -- for instances, wrt how data
                            is gathered, processed and classified, how
                            hypotheses are operationalized, etc., to
                            make sure that the results obtained by
                            somebody are also accepted by others (just
                            think of the 5%-threshold for statistical
                            significance, which is just a matter of
                            convention).<o:p></o:p></p>
                          <p class="MsoNormal"
                            style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">I'm
                            aware that this type of remark is annoying
                            for some of you. I teach both corpus
                            linguistics and typology. In corpus
                            linguistics our students deal with very
                            basic questions of empirical research --
                            like the traditional 'quality criteria' --
                            e.g. (external, internal) validity,
                            objectivity, reliability -- and then, in
                            typology, we read important publications by
                            major authorities of the field where these
                            criteria are simply not applied, sometimes
                            the statistics are faulty, and students do
                            enquire about this. What can I say? There
                            are no research standards in typology? There
                            is an ongoing discussion about
                            "arbitrary/subjective/random/disjunctive
                            comparative concepts" on the Lingtype-list?
                            I'm afraid it wouldn't convince them. What I
                            say is that typology still has some way to
                            go to in terms of research methods. There
                            are many non-trivial problems, as we have
                            seen in various discussions on this list,
                            and we should be aware that linguistic data
                            is sui generis (for instance, I think we
                            can't adopt just any method/software package
                            from genetics). But we shouldn't use
                            "authority" as a criterion in our
                            methodological choices, and the choices
                            shouldn't be made in such a way to
                            legitimize our own research 'ex post'.<o:p></o:p></p>
                          <p class="MsoNormal"
                            style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">Volker<o:p></o:p></p>
                          <p class="MsoNormal"><o:p> </o:p></p>
                          <div>
                            <p class="MsoNormal">Am 18.11.2017 um 07:36
                              schrieb Dryer, Matthew:<o:p></o:p></p>
                          </div>
                          <blockquote
                            style="margin-top:5.0pt;margin-bottom:5.0pt">
                            <div>
                              <p class="MsoNormal"
                                style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">With
                                respect to Martin’s comment<o:p></o:p></p>
                              <div>
                                <p class="MsoNormal"> <o:p></o:p></p>
                              </div>
                              <p class="MsoNormal"
                                style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">“<span
                                  style="font-size:14.0pt">It is my
                                  impression that such ortho-affixes (=
                                  forms written as affixes) are perhaps
                                  even more common than “phonologically
                                  weak” ortho-affixes, but this is an
                                  empirical question (in his 2015 ALT
                                  abstract, Matthew mentions 248
                                  languages with weak affixes, but 308
                                  languages with only affixes of the
                                  Tauya type, apparently confirming my
                                  impression).</span>”<o:p></o:p></p>
                              <div>
                                <p class="MsoNormal"> <o:p></o:p></p>
                              </div>
                              <p class="MsoNormal"
                                style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">I
                                realize that this is a reasonable
                                inference from my abstract, but one
                                often has to simplify things for the
                                purposes of an abstract. My definition
                                of a weak affix is very narrow and many
                                if not most affixes that are not weak
                                affixes by my narrow criteria can still
                                be shown to be attached phonologically
                                by broader criteria. Furthermore, I also
                                treat a morpheme as an affix for the
                                purposes of this study if it triggers
                                phonologically conditioned allomorphy in
                                stems it attaches to and it is clear
                                from Macdonald’s description of Tauya
                                that some of the ortho-affixes Martin
                                mentions do trigger phonologically
                                conditioned allomorphy in stems they
                                attach to (pp 54, 72, 74, 79).
                                <o:p></o:p></p>
                              <div>
                                <p class="MsoNormal"> <o:p></o:p></p>
                              </div>
                              <p class="MsoNormal"
                                style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">I
                                counted an affix as weak for the
                                purposes of the study in my 2015 ALT
                                talk only if the description of it in a
                                grammar makes clear that it is
                                nonsyllabic (or has nonsyllabic
                                allomorphs) or that it exhibits
                                phonologically allomorphy or triggers
                                phonologically conditioned allomorphy in
                                adjacent stems. But in many grammars, it
                                is only in the discussion of phonology
                                that it becomes clear that a given affix
                                exhibits phonologically conditioned
                                allomorphy or that it triggers
                                phonologically conditioned allomorphy in
                                adjacent stems. But because I wanted to
                                include a large sample of languages and
                                because it is often unclear from
                                discussions of phonology whether
                                particular rules apply to particular
                                affixes or stems such affixes combine
                                with, I adopted the procedure of not
                                consulting the discussions of phonology
                                in classifying ortho-affixes as weak.
                                This made sense for my 2015 ALT talk
                                since I was examining whether there is a
                                suffixing preference and restricting
                                attention to weak affixes so defined
                                applies equally to prefixes and
                                suffixes. For a different type of
                                typological study, this would have been
                                inappropriate. This illustrates how
                                comparative concepts are specific to
                                particular typological studies.<o:p></o:p></p>
                              <div>
                                <p class="MsoNormal"> <o:p></o:p></p>
                              </div>
                              <p class="MsoNormal"
                                style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">Furthermore,
                                there are other factors that I did not
                                examine that are relevant to whether a
                                given ortho-affix is attached
                                phonologically. There may be clear
                                evidence from allophonic rules, but it
                                is often very unclear from grammatical
                                descriptions whether particular
                                allophonic rules apply to particular
                                ortho-affixes or stems to which
                                ortho-affixes are attached. And even if
                                the information is there in the
                                grammatical description, it may take a
                                lot of work to see whether they apply to
                                a particular affix. For example, careful
                                examination of Macdonald’s description
                                of Tauya implies that the benefactive
                                ortho-affix
                                <i>-pe</i> that Martin mentions is
                                attached phonologically, since she gives
                                examples of phonetic representations of
                                forms containing this morpheme where it
                                takes the form [-be] after /m/ ([tembe]
                                on page 54).<o:p></o:p></p>
                              <div>
                                <p class="MsoNormal"> <o:p></o:p></p>
                              </div>
                              <p class="MsoNormal"
                                style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">There
                                might also be evidence from stress, but
                                still be unclear how stress is assigned
                                to forms including ortho-affixes. For
                                example, Tauya has word-final stress,
                                but it is not clear from Macdonald’s
                                description whether this means that
                                nouns bearing the ortho-affixes that
                                Martin mentions take stress on the
                                ortho-affix.<o:p></o:p></p>
                              <div>
                                <p class="MsoNormal"> <o:p></o:p></p>
                              </div>
                              <p class="MsoNormal"
                                style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">Some
                                of you may have noticed that what I say
                                here contradicts what I said in my
                                earlier email about comparative concepts
                                needing to be exhaustive. The
                                comparative concept I used in my 2015
                                ALT talk was not exhaustive and was in
                                fact disjunctive. Since that seemed
                                appropriate for that study, this
                                suggests that one cannot rule out
                                disjunctively defined comparative
                                concepts. I sympathize with Martin’s
                                objecting to disjunctive comparative
                                concepts as a way to continue to use
                                confusing and ambiguous terms and I
                                agree that there is something odd about
                                arbitrary disjunctive comparative
                                concepts, but it is a mistake to simply
                                rule out disjunctive comparative
                                concepts.<o:p></o:p></p>
                              <div>
                                <p class="MsoNormal"> <o:p></o:p></p>
                              </div>
                              <p class="MsoNormal"
                                style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">I
                                should note finally that while it is
                                clear that the ortho-affixes that Martin
                                mentions are attached phonologically,
                                they are actually not affixes by either
                                his criteria or mine since they are
                                clitics that attach to postnominal
                                modifiers. [Martin has written about
                                problems with the use of the term
                                “clitic”. I am in complete agreement
                                with him about this. But I use the term
                                here and elsewhere in my research
                                (including my upcoming ALT talk on the
                                encliticization preference) as a label
                                for a comparative concept for
                                grammatical morphemes that are
                                phonologically attached but attach to
                                stems of more than one stem class.]<o:p></o:p></p>
                              <div>
                                <p class="MsoNormal"> <o:p></o:p></p>
                              </div>
                              <p class="MsoNormal"
                                style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">Matthew<o:p></o:p></p>
                            </div>
                            <div>
                              <p class="MsoNormal"><o:p> </o:p></p>
                            </div>
                            <div style="border:none;border-top:solid
                              #B5C4DF 1.0pt;padding:3.0pt 0cm 0cm 0cm">
                              <p class="MsoNormal"><b>From: </b>Lingtyp
                                <<a moz-do-not-send="true"
                                  href="mailto:lingtyp-bounces@listserv.linguistlist.org"
                                  target="_blank">lingtyp-bounces@listserv.linguistlist.org</a>>
                                on behalf of Martin Haspelmath <<a
                                  moz-do-not-send="true"
                                  href="mailto:haspelmath@shh.mpg.de"
                                  target="_blank">haspelmath@shh.mpg.de</a>><br>
                                <b>Date: </b>Thursday, November 16,
                                2017 at 7:14 PM<br>
                                <b>To: </b>"<a moz-do-not-send="true"
                                  href="mailto:lingtyp@listserv.linguistlist.org"
                                  target="_blank">lingtyp@listserv.linguistlist.org</a>"
                                <<a moz-do-not-send="true"
                                  href="mailto:lingtyp@listserv.linguistlist.org"
                                  target="_blank">lingtyp@listserv.linguistlist.org</a>><br>
                                <b>Subject: </b>Re: [Lingtyp] wordhood:
                                bonded vs. bound<o:p></o:p></p>
                            </div>
                            <div>
                              <p class="MsoNormal"><o:p> </o:p></p>
                            </div>
                            <div>
                              <div>
                                <p class="MsoNormal">Matthew Dryer
                                  thinks that wordhood is generally
                                  understood by grammar authors in terms
                                  of
                                  <b>bondedness</b> (= phonological
                                  weakness, as shown by nonsyllabicity
                                  and phono-conditioned allomorphy), not
                                  in terms of
                                  <b>boundness</b> (= inability to occur
                                  in isolation). <o:p></o:p></p>
                                <div>
                                  <p class="MsoNormal"> <o:p></o:p></p>
                                </div>
                                <p class="MsoNormal">I don’t know if
                                  this is true, but Matthew actually
                                  recognizes that grammars often
                                  describe grammatical markers as
                                  “affixes” even when they do not show
                                  the two “phonological weakness” (or
                                  bondedness) features.
                                  <o:p></o:p></p>
                                <div>
                                  <p class="MsoNormal"> <o:p></o:p></p>
                                </div>
                                <p class="MsoNormal">For example, Tauya
                                  (a language of New Guinea) is said to
                                  have (syllabic) case suffixes, but
                                  these never show any allomorphy, e.g.
                                  <o:p></o:p></p>
                                <div>
                                  <p class="MsoNormal"> <o:p></o:p></p>
                                </div>
                                <p class="MsoNormal">fena’a-ni
                                  [woman-ERG]<br>
                                  na-pe [you-BEN]<br>
                                  wate-’usa [house-INESS]<br>
                                  Aresa-nani [Aresa-ALL]<br>
                                  Tauya-sami [Tauya-ABL] (MacDonald
                                  1990: 119-126) <o:p></o:p></p>
                                <div>
                                  <p class="MsoNormal"> <o:p></o:p></p>
                                </div>
                                <p class="MsoNormal">It is my impression
                                  that such ortho-affixes (= forms
                                  written as affixes) are perhaps even
                                  more common than “phonologically weak”
                                  ortho-affixes, but this is an
                                  empirical question (in his 2015 ALT
                                  abstract, Matthew mentions 248
                                  languages with weak affixes, but 308
                                  languages with only affixes of the
                                  Tauya type, apparently confirming my
                                  impression).
                                  <o:p></o:p></p>
                                <div>
                                  <p class="MsoNormal"> <o:p></o:p></p>
                                </div>
                                <p class="MsoNormal">For this reason, I
                                  have suggested that the stereotypical
                                  “affix” notion should perhaps be
                                  captured in terms of boundness
                                  together with single-root-class
                                  adjacency. Since the Tauya
                                  case-markers attach only to nouns,
                                  they count as affixes; by contrast, if
                                  a bound role marker attaches to both
                                  nouns (English “for children”) and
                                  adjectives (“for older children”) as
                                  well as to other elements (“for many
                                  children”), we do not regard it as an
                                  affix (but as a preposition), even if
                                  it is bound (= does not occur in
                                  isolation; English "for" does not). <o:p></o:p></p>
                                <div>
                                  <p class="MsoNormal"> <o:p></o:p></p>
                                </div>
                                <p class="MsoNormal">Matthew quite
                                  rightly points out that this notion of
                                  boundness (which goes back at least to
                                  Bloomfield 1933: §10.1) implies that
                                  most function words in English are
                                  bound, and in fact most function words
                                  in most languages are bound – but this
                                  is exactly what we want, I feel,
                                  because the best way to define a
                                  “function word” is as a bound element
                                  that is not an affix. Linguists often
                                  think of function words (or
                                  “functional categories”) as defined
                                  semantically, but it is actually very
                                  hard to say what is the
                                  semantic(-pragmatic) difference
                                  between a plural marker and a word
                                  like “several”, between a dual marker
                                  and the word “two”, between a
                                  past-tense marker and the expression
                                  “in the past”, or between a comitative
                                  marker and the word “accompany”. It
                                  seems to me that these distinctions
                                  are best characterized in terms of
                                  boundness, i.e. inability to occur in
                                  isolation.
                                  <o:p></o:p></p>
                                <div>
                                  <p class="MsoNormal"> <o:p></o:p></p>
                                </div>
                                <p class="MsoNormal">It may be true that
                                  occurrence in isolation is a feature
                                  of an element that is not easy to
                                  elicit from speakers, but in actual
                                  language use, there are a very large
                                  number of very short utterances, so at
                                  least positive evidence for free
                                  status (=non-bound status) is not
                                  difficult to obtain. <o:p></o:p></p>
                                <div>
                                  <p class="MsoNormal"> <o:p></o:p></p>
                                </div>
                                <p class="MsoNormal">In any event, it
                                  seems clear to me that some key
                                  concepts of grammatical typology such
                                  as “flag” (= bound role marker on a
                                  nominal) and “person index” (= bound
                                  person marker, generally on a verb)
                                  require the Bloomfieldian boundness
                                  notion, and that these concepts are
                                  much easier to work with in typology
                                  than the traditional stereotypical
                                  notions of “case”, “adposition”,
                                  “agreement marker”, and “pronominal
                                  clitic”. (For bound person forms, this
                                  was a major lesson of Anna
                                  Siewierska’s 2004 book “Person”.) <o:p></o:p></p>
                                <div>
                                  <p class="MsoNormal"> <o:p></o:p></p>
                                </div>
                                <p class="MsoNormal"
                                  style="margin-bottom:12.0pt">Best,<br>
                                  Martin<o:p></o:p></p>
                                <div>
                                  <p class="MsoNormal">On 14.11.17
                                    07:02, Dryer, Matthew wrote:<o:p></o:p></p>
                                </div>
                                <blockquote
                                  style="margin-top:5.0pt;margin-bottom:5.0pt">
                                  <div>
                                    <p class="MsoNormal"
                                      style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                        style="font-size:10.5pt">I have
                                        a number of problems with
                                        Martin’s proposal:<o:p></o:p></span></p>
                                    <div>
                                      <p class="MsoNormal"><span
                                          style="font-size:10.5pt"> <o:p></o:p></span></p>
                                    </div>
                                    <p class="MsoNormal"
                                      style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                        style="font-size:14.0pt">"<b>Here’s
                                          a proposal for defining a
                                          notion of “affix”, in such a
                                          way that the results do not go
                                          too much against our
                                          intuitions or stereotypes:</b></span><span
                                        style="font-size:10.5pt"><o:p></o:p></span></p>
                                    <div>
                                      <p class="MsoNormal"><span
                                          style="font-size:14.0pt"> </span><span
                                          style="font-size:10.5pt"><o:p></o:p></span></p>
                                    </div>
                                    <p class="MsoNormal"
                                      style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><b><span
                                          style="font-size:14.0pt">An
                                          affix is a bound form that
                                          always occurs together with a
                                          root of the same root-class
                                          and is never separated from
                                          the root by a free form or a
                                          non-affixal bound form."</span></b><span
                                        style="font-size:10.5pt"><o:p></o:p></span></p>
                                    <div>
                                      <p class="MsoNormal"><span
                                          style="font-size:10.5pt"> <o:p></o:p></span></p>
                                    </div>
                                    <p class="MsoNormal"
                                      style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                        style="font-size:10.5pt">If one
                                        examines the notion of “bound”
                                        from his 2013 paper, I believe
                                        it implies a comparative concept
                                        of affix that differs greatly
                                        from what most linguists (at
                                        last most non-generative
                                        linguists) understand by the
                                        term. That’s not a problem for
                                        it as a comparative concept, but
                                        it is a comparative concept that
                                        differs considerably from the
                                        stereotype.<o:p></o:p></span></p>
                                    <div>
                                      <p class="MsoNormal"><span
                                          style="font-size:10.5pt"> <o:p></o:p></span></p>
                                    </div>
                                    <p class="MsoNormal"
                                      style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                        style="font-size:10.5pt">Martin’s
                                        definition of “free and “bound”
                                        from his 2013 paper is as
                                        follows:<o:p></o:p></span></p>
                                    <div>
                                      <p class="MsoNormal"><span
                                          style="font-size:10.5pt"> <o:p></o:p></span></p>
                                    </div>
                                    <p class="MsoNormal"
                                      style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><b><span
style="font-size:13.5pt;font-family:"Times",serif">"But
                                          distinguishing in a general
                                          way between bound elements and
                                          free elements is quite
                                          straightforward, because there
                                          is a single criterion: Free
                                          forms are forms that can occur
                                          on their own, i.e. in a
                                          complete (possibly elliptical)
                                          utterance (Bloomfield 1933:
                                          160). This criterion
                                          correlates very highly with
                                          the criterion of contrastive
                                          use: Only free forms can be
                                          used contrastively."</span></b><o:p></o:p></p>
                                    <div>
                                      <p class="MsoNormal"><span
                                          style="font-size:13.0pt;font-family:"Times",serif"> </span><span
                                          style="font-size:10.5pt"><o:p></o:p></span></p>
                                    </div>
                                    <p class="MsoNormal"
                                      style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
style="font-size:10.5pt;font-family:"Times",serif">First, I
                                        find the notion of complete
                                        utterance ambiguous. Does it
                                        mean utterances in normal speech
                                        or does it include
                                        metalinguistic uses (like “What
                                        is the last word in the sentence
                                        “Who are you going with”? Answer
                                        “with”). I would assume that it
                                        does not include such
                                        metalinguistic uses. But then
                                        many if not most so-called
                                        function words in English would
                                        count as bound since they cannot
                                        be used as complete utterances.
                                        Perhaps other speakers of
                                        English would have different
                                        intuitions, but if so that only
                                        indicates the lack of clarity in
                                        the notion. Furthermore, for
                                        many function words in English,
                                        I am not sure how to judge
                                        whether they can occur alone as
                                        utterances. Many such so-called
                                        function words would appear to
                                        count as bound by Martin’s
                                        definition, though they would
                                        not count as affixes since they
                                        lack other properties in his
                                        definition of “affix”.</span><span
                                        style="font-size:10.5pt"><o:p></o:p></span></p>
                                    <div>
                                      <p class="MsoNormal"><span
                                          style="font-size:10.5pt;font-family:"Times",serif"> </span><span
                                          style="font-size:10.5pt"><o:p></o:p></span></p>
                                    </div>
                                    <p class="MsoNormal"
                                      style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
style="font-size:10.5pt;font-family:"Times",serif">Second,
                                        many languages have grammatical
                                        morphemes that must occur
                                        adjacent to an open class word
                                        but which behave as separate
                                        words phonologically. These
                                        would all apparently count as
                                        affixes by Martin’s definition.
                                        Again, I have no problem with
                                        this as a comparative concept,
                                        only that it means his notion of
                                        affix deviates considerably from
                                        the stereotype.</span><span
                                        style="font-size:10.5pt"><o:p></o:p></span></p>
                                    <div>
                                      <p class="MsoNormal"><span
                                          style="font-size:10.5pt;font-family:"Times",serif"> </span><span
                                          style="font-size:10.5pt"><o:p></o:p></span></p>
                                    </div>
                                    <p class="MsoNormal"
                                      style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
style="font-size:10.5pt;font-family:"Times",serif">Third,
                                        Martin says that his criterion
                                        “correlates very highly with the
                                        criterion of contrastive use”.
                                        But by my intuitions, the
                                        ability to occur as complete
                                        utterances does not correlate
                                        closely with the criterion of
                                        contrastive use, since most
                                        so-called function words CAN
                                        occur with contrastive use (such
                                        as can in this sentence!), as
                                        can some morphemes that are
                                        conventionally treated as
                                        affixes, like <i>un-</i> in
                                        “I’m not happy, I’m UNhappy”. Of
                                        course, Martin might argue that
                                        <i>un-</i> is more like
                                        so-called function words and
                                        less like morphemes
                                        conventionally treated as
                                        affixes. But the fact remains
                                        that
                                        <i>un-</i> is easily the locus
                                        of contrast but cannot be used
                                        as a complete utterance. I thus
                                        see no evidence of a close
                                        correlation between the ability
                                        to occur as a complete utterance
                                        and the ability to be the locus
                                        of contrast.</span><span
                                        style="font-size:10.5pt"><o:p></o:p></span></p>
                                    <div>
                                      <p class="MsoNormal"><span
                                          style="font-size:10.5pt;font-family:"Times",serif"> </span><span
                                          style="font-size:10.5pt"><o:p></o:p></span></p>
                                    </div>
                                    <p class="MsoNormal"
                                      style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
style="font-size:10.5pt;font-family:"Times",serif">Finally, it
                                        is my experience that languages
                                        differ in their conventions
                                        regarding what can be a complete
                                        utterance. Imagine two closely
                                        related languages that differ in
                                        their grammatical rules
                                        governing what is a complete
                                        utterance. By Martin’s
                                        definition, there might be a
                                        large number of morphemes that
                                        count as separate words in one
                                        language but as affixes in the
                                        other language. This strikes me
                                        as odd. It seems odd to have a
                                        criterion for what is a word and
                                        what is an affix so dependent on
                                        the grammatical rules in the
                                        language for what constitutes a
                                        complete utterance.</span><span
                                        style="font-size:10.5pt"><o:p></o:p></span></p>
                                    <div>
                                      <p class="MsoNormal"><span
                                          style="font-size:10.5pt;font-family:"Times",serif"> </span><span
                                          style="font-size:10.5pt"><o:p></o:p></span></p>
                                    </div>
                                    <p class="MsoNormal"
                                      style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
style="font-size:10.5pt;font-family:"Times",serif">Matthew</span><span
                                        style="font-size:10.5pt"><o:p></o:p></span></p>
                                  </div>
                                  <div>
                                    <p class="MsoNormal"><span
                                        style="font-size:10.5pt"><o:p> </o:p></span></p>
                                  </div>
                                  <div
                                    style="border:none;border-top:solid
                                    #B5C4DF 1.0pt;padding:3.0pt 0cm 0cm
                                    0cm">
                                    <p class="MsoNormal"><b>From: </b>Lingtyp
                                      <<a moz-do-not-send="true"
                                        href="mailto:lingtyp-bounces@listserv.linguistlist.org"
                                        target="_blank">lingtyp-bounces@listserv.linguistlist.org</a>>
                                      on behalf of Martin Haspelmath
                                      <<a moz-do-not-send="true"
                                        href="mailto:haspelmath@shh.mpg.de"
                                        target="_blank">haspelmath@shh.mpg.de</a>><br>
                                      <b>Date: </b>Sunday, November 12,
                                      2017 at 10:47 PM<br>
                                      <b>To: </b>"<a
                                        moz-do-not-send="true"
                                        href="mailto:lingtyp@listserv.linguistlist.org"
                                        target="_blank">lingtyp@listserv.linguistlist.org</a>"
                                      <<a moz-do-not-send="true"
                                        href="mailto:lingtyp@listserv.linguistlist.org"
                                        target="_blank">lingtyp@listserv.linguistlist.org</a>><br>
                                      <b>Subject: </b>Re: [Lingtyp]
                                      wordhood<o:p></o:p></p>
                                  </div>
                                  <div>
                                    <p class="MsoNormal"><span
                                        style="font-size:10.5pt"><o:p> </o:p></span></p>
                                  </div>
                                  <div>
                                    <div>
                                      <p class="MsoNormal"
                                        style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                          style="font-size:10.5pt">Mattis
                                          List and Balthasar Bickel
                                          rightly emphasize that “word”
                                          is not a Platonic entity (a
                                          natural kind) that exists in
                                          advance of language learning
                                          or linguistic analysis – few
                                          linguists would disagree here,
                                          not even generativists (who
                                          otherwise liberally assume
                                          natural-kind catgeories).<o:p></o:p></span></p>
                                      <div>
                                        <p class="MsoNormal"><span
                                            style="font-size:10.5pt"> <o:p></o:p></span></p>
                                      </div>
                                      <p class="MsoNormal"
                                        style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                          style="font-size:10.5pt">But I
                                          think many linguists still ACT
                                          AS IF there were such a
                                          natural kind, because the
                                          “word” notion is a crucial
                                          ingredient to a number of
                                          other notions that linguists
                                          use routinely – e.g. “gender”,
                                          which is typically defined in
                                          terms of “agreement” (which is
                                          defined in terms of
                                          inflectional marking on
                                          targets; and inflection is
                                          defined in terms of “word”).<o:p></o:p></span></p>
                                      <div>
                                        <p class="MsoNormal"><span
                                            style="font-size:10.5pt"> <o:p></o:p></span></p>
                                      </div>
                                      <p class="MsoNormal"
                                        style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                          style="font-size:10.5pt">So is
                                          it possible to define a
                                          comparative concept ‘word’
                                          that applies to all languages
                                          equally, and that accords
                                          reasonably with our
                                          stereotypes? Note that I
                                          didn’t deny this in my 2011
                                          paper, I just said that nobody
                                          had come up with a
                                          satisfactory definition (that
                                          could be used, for instance,
                                          in defining “gender” or
                                          “polysynthesis”). So I’ll be
                                          happy to contribute to a
                                          discussion on how to make
                                          progress on defining “word”.<o:p></o:p></span></p>
                                      <div>
                                        <p class="MsoNormal"><span
                                            style="font-size:10.5pt"> <o:p></o:p></span></p>
                                      </div>
                                      <p class="MsoNormal"
                                        style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                          style="font-size:10.5pt">Larry
                                          Hyman notes that other notions
                                          like “syllable” and “sentence”
                                          are also problematic in that
                                          they also “leak”. However, I
                                          think it is important to
                                          distinguish two situations of
                                          “slipperiness”:<o:p></o:p></span></p>
                                      <div>
                                        <p class="MsoNormal"><span
                                            style="font-size:10.5pt"> <o:p></o:p></span></p>
                                      </div>
                                      <p class="MsoNormal"
                                        style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                          style="font-size:10.5pt">(1)
                                          “Leakage” of definitions due
                                          to vague defining notions<o:p></o:p></span></p>
                                      <div>
                                        <p class="MsoNormal"><span
                                            style="font-size:10.5pt"> <o:p></o:p></span></p>
                                      </div>
                                      <p class="MsoNormal"
                                        style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                          style="font-size:10.5pt">(2)
                                          Incoherence of definitions due
                                          to the use of different
                                          criteria in different
                                          languages<o:p></o:p></span></p>
                                      <div>
                                        <p class="MsoNormal"><span
                                            style="font-size:10.5pt"> <o:p></o:p></span></p>
                                      </div>
                                      <p class="MsoNormal"
                                        style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                          style="font-size:10.5pt">The
                                          first can be addressed by
                                          tightening the defining
                                          notions, but the second is
                                          fatal.<o:p></o:p></span></p>
                                      <div>
                                        <p class="MsoNormal"><span
                                            style="font-size:10.5pt"> <o:p></o:p></span></p>
                                      </div>
                                      <p class="MsoNormal"
                                        style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                          style="font-size:10.5pt">To
                                          take up Östen Dahl’s example
                                          of the “family” notion: In one
                                          culture, a family might be
                                          said to be a set of minimally
                                          three living people consisting
                                          of two adults (regardless of
                                          gender) living in a romantic
                                          relationship plus all their
                                          descendants. In another
                                          culture, a family might be
                                          defined as a married couple
                                          consisting of a man and a
                                          woman plus all their living
                                          direct ancestors, all their
                                          (great) uncles and (great)
                                          aunts, and all the descendants
                                          of all of these.<o:p></o:p></span></p>
                                      <div>
                                        <p class="MsoNormal"><span
                                            style="font-size:10.5pt"> <o:p></o:p></span></p>
                                      </div>
                                      <p class="MsoNormal"
                                        style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                          style="font-size:10.5pt">With
                                          two family concepts as
                                          different as these, it is
                                          obviously not very interesting
                                          to ask general cross-cultural
                                          questions about “families”
                                          (e.g. “How often do all family
                                          members have meals
                                          together?”). So the use of
                                          different criteria for
                                          different cultures is fatal
                                          here.<o:p></o:p></span></p>
                                      <div>
                                        <p class="MsoNormal"><span
                                            style="font-size:10.5pt"> <o:p></o:p></span></p>
                                      </div>
                                      <p class="MsoNormal"
                                        style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                          style="font-size:10.5pt">What
                                          I find worrying is that
                                          linguists often seem to accept
                                          incoherent definitions of
                                          comparative concepts (this was
                                          emphasized especially in my
                                          2015 paper on defining vs.
                                          diagnosing categories).
                                          Different diagnostics in
                                          different languages would not
                                          be fatal if “word” were a
                                          Platonic (natural-kind)
                                          concept, but if we are not
                                          born with a “word” category,
                                          typologists need to use the
                                          SAME criteria for all
                                          languages.<o:p></o:p></span></p>
                                      <div>
                                        <p class="MsoNormal"><span
                                            style="font-size:10.5pt"> <o:p></o:p></span></p>
                                      </div>
                                      <p class="MsoNormal"
                                        style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                          style="font-size:10.5pt">So
                                          here’s a proposal for defining
                                          a notion of “simple
                                          morphosyntactic word”:<o:p></o:p></span></p>
                                      <div>
                                        <p class="MsoNormal"><span
                                            style="font-size:10.5pt"> <o:p></o:p></span></p>
                                      </div>
                                      <p class="MsoNormal"
                                        style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><b><span
                                            style="font-size:10.5pt">A
                                            simple morphosyntactic word
                                            is a form that consists of
                                            (minimally) a root, plus any
                                            affixes.</span></b><span
                                          style="font-size:10.5pt"><o:p></o:p></span></p>
                                      <div>
                                        <p class="MsoNormal"><span
                                            style="font-size:10.5pt"> <o:p></o:p></span></p>
                                      </div>
                                      <p class="MsoNormal"
                                        style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                          style="font-size:10.5pt">Here’s
                                          a proposal for defining a
                                          notion of “affix”, in such a
                                          way that the results do not go
                                          too much against our
                                          intuitions or stereotypes:<o:p></o:p></span></p>
                                      <div>
                                        <p class="MsoNormal"><span
                                            style="font-size:10.5pt"> <o:p></o:p></span></p>
                                      </div>
                                      <p class="MsoNormal"
                                        style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><b><span
                                            style="font-size:10.5pt">An
                                            affix is a bound form that
                                            always occurs together with
                                            a root of the same
                                            root-class and is never
                                            separated from the root by a
                                            free form or a non-affixal
                                            bound form.</span></b><span
                                          style="font-size:10.5pt"><o:p></o:p></span></p>
                                      <div>
                                        <p class="MsoNormal"><span
                                            style="font-size:10.5pt"> <o:p></o:p></span></p>
                                      </div>
                                      <p class="MsoNormal"
                                        style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                          style="font-size:10.5pt">These
                                          definitions make use of the
                                          notions of “root” and
                                          “root-class” (defined in
                                          Haspelmath 2012) and  “bound
                                          (form)” vs. “free (form)”
                                          (defined in Haspelmath 2013).
                                          All these show leakage as in
                                          (1) above, but they are
                                          equally applicable to all
                                          languages, so they are not
                                          incoherent. (I thank Harald
                                          Hammarström for a helpful
                                          discussion that helped me to
                                          come up with the above
                                          definitions, which I had not
                                          envisaged in 2011.)<o:p></o:p></span></p>
                                      <div>
                                        <p class="MsoNormal"><span
                                            style="font-size:10.5pt"> <o:p></o:p></span></p>
                                      </div>
                                      <p class="MsoNormal"
                                        style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                          style="font-size:10.5pt">(What
                                          I don’t know at the moment is
                                          how to relate “simple
                                          morphosyntactic word” to
                                          “morphosyntactic word” in
                                          general, because I cannot
                                          distinguish compounds from
                                          phrases comparatively; and I
                                          don’t know what to do with
                                          “phonological word”.)<o:p></o:p></span></p>
                                      <div>
                                        <p class="MsoNormal"><span
                                            style="font-size:10.5pt"> <o:p></o:p></span></p>
                                      </div>
                                      <p class="MsoNormal"
                                        style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                          style="font-size:10.5pt">Crucially,
                                          the definitions above make use
                                          of a number of basic concepts
                                          that apply to ALL languages in
                                          the SAME way. David Gil’s
                                          proposal, to measure “bond
                                          strength” by means of a range
                                          of language-particular
                                          phenomena, falls short of this
                                          requirement (as already hinted
                                          by Eitan Grossman). Note that
                                          the problem I have with
                                          David’s proposal is not that
                                          it provides no categorical
                                          contrasts (recall my
                                          acceptance of vagueness in (1)
                                          above), but that there is no
                                          way of telling which phenomena
                                          should count as measuring bond
                                          strength.<o:p></o:p></span></p>
                                      <div>
                                        <p class="MsoNormal"><span
                                            style="font-size:10.5pt"> <o:p></o:p></span></p>
                                      </div>
                                      <p class="MsoNormal"
                                        style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                          style="font-size:10.5pt">David’s
                                          approach resembles Keenan’s
                                          (1976) attempt at defining
                                          “subject” (perhaps not by
                                          accident, because Ed Keenan
                                          was David’s PhD supervisor),
                                          but I have a similar objection
                                          to Keenan: If different
                                          criteria are used for
                                          different languages, how do we
                                          know that we are measuring the
                                          same phenomenon across
                                          languages? Measuring X by
                                          means of Y makes sense only if
                                          we know independently that X
                                          and Y are very highly
                                          correlated. But do we know
                                          this, for subjects, or for
                                          bond strength?<o:p></o:p></span></p>
                                      <div>
                                        <p class="MsoNormal"><span
                                            style="font-size:10.5pt"> <o:p></o:p></span></p>
                                      </div>
                                      <p class="MsoNormal"
                                        style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                          style="font-size:10.5pt">Best,<o:p></o:p></span></p>
                                      <p class="MsoNormal"
                                        style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"><span
                                          style="font-size:10.5pt">Martin<o:p></o:p></span></p>
                                      <div>
                                        <p class="MsoNormal"><span
                                            style="font-size:10.5pt"> <o:p></o:p></span></p>
                                      </div>
                                      <p class="MsoNormal"><span
                                          style="font-size:10.5pt"><br>
                                          <br>
                                          <o:p></o:p></span></p>
                                      <pre>-- <o:p></o:p></pre>
                                      <pre>Martin Haspelmath (<a moz-do-not-send="true" href="mailto:haspelmath@shh.mpg.de" target="_blank">haspelmath@shh.mpg.de</a>)<o:p></o:p></pre>
                                      <pre>Max Planck Institute for the Science of Human History<o:p></o:p></pre>
                                      <pre>Kahlaische Strasse 10       <o:p></o:p></pre>
                                      <pre>D-07745 Jena  <o:p></o:p></pre>
                                      <pre>&<o:p></o:p></pre>
                                      <pre>Leipzig University <o:p></o:p></pre>
                                      <pre>IPF 141199<o:p></o:p></pre>
                                      <pre>Nikolaistrasse 6-10<o:p></o:p></pre>
                                      <pre>D-04109 Leipzig    <o:p></o:p></pre>
                                      <pre><o:p> </o:p></pre>
                                      <pre><o:p> </o:p></pre>
                                      <pre><o:p> </o:p></pre>
                                      <pre><o:p> </o:p></pre>
                                      <pre><o:p> </o:p></pre>
                                    </div>
                                  </div>
                                </blockquote>
                                <p class="MsoNormal"><br>
                                  <br>
                                  <o:p></o:p></p>
                                <pre>-- <o:p></o:p></pre>
                                <pre>Martin Haspelmath (<a moz-do-not-send="true" href="mailto:haspelmath@shh.mpg.de" target="_blank">haspelmath@shh.mpg.de</a>)<o:p></o:p></pre>
                                <pre>Max Planck Institute for the Science of Human History<o:p></o:p></pre>
                                <pre>Kahlaische Strasse 10     <o:p></o:p></pre>
                                <pre>D-07745 Jena  <o:p></o:p></pre>
                                <pre>&<o:p></o:p></pre>
                                <pre>Leipzig University <o:p></o:p></pre>
                                <pre>IPF 141199<o:p></o:p></pre>
                                <pre>Nikolaistrasse 6-10<o:p></o:p></pre>
                                <pre>D-04109 Leipzig    <o:p></o:p></pre>
                                <pre><o:p> </o:p></pre>
                                <pre><o:p> </o:p></pre>
                                <pre><o:p> </o:p></pre>
                                <pre><o:p> </o:p></pre>
                                <pre><o:p> </o:p></pre>
                              </div>
                            </div>
                            <p class="MsoNormal"><br>
                              <br>
                              <o:p></o:p></p>
                            <pre>_______________________________________________<o:p></o:p></pre>
                            <pre>Lingtyp mailing list<o:p></o:p></pre>
                            <pre><a moz-do-not-send="true" href="mailto:Lingtyp@listserv.linguistlist.org" target="_blank">Lingtyp@listserv.linguistlist.org</a><o:p></o:p></pre>
                            <pre><a moz-do-not-send="true" href="http://listserv.linguistlist.org/mailman/listinfo/lingtyp" target="_blank">http://listserv.linguistlist.org/mailman/listinfo/lingtyp</a><span class="hoenzb"><span style="color:#888888"><o:p></o:p></span></span></pre>
                          </blockquote>
                          <p class="MsoNormal"><span
                              style="color:#888888"><br>
                              <br>
                            </span><span class="hoenzb"><span
                                style="color:#888888"><o:p></o:p></span></span></p>
                          <pre><span style="color:#888888">-- <o:p></o:p></span></pre>
                          <pre><span style="color:#888888">Prof. Volker Gast<o:p></o:p></span></pre>
                          <pre><span style="color:#888888">English and American Studies<o:p></o:p></span></pre>
                          <pre><span style="color:#888888">Ernst-Abbe-PLatz 8<o:p></o:p></span></pre>
                          <pre><span style="color:#888888">D-07743 Jena<o:p></o:p></span></pre>
                          <pre><span style="color:#888888"><o:p> </o:p></span></pre>
                          <pre><span style="color:#888888">Fon: ++49 3641 9-44546<o:p></o:p></span></pre>
                          <pre><span style="color:#888888">Fax: ++49 3641 9-44542</span><o:p></o:p></pre>
                        </div>
                        <p class="MsoNormal"
                          style="margin-bottom:12.0pt"><br>
_______________________________________________<br>
                          Lingtyp mailing list<br>
                          <a moz-do-not-send="true"
                            href="mailto:Lingtyp@listserv.linguistlist.org">Lingtyp@listserv.linguistlist.org</a><br>
                          <a moz-do-not-send="true"
                            href="http://listserv.linguistlist.org/mailman/listinfo/lingtyp"
                            target="_blank">http://listserv.linguistlist.org/mailman/listinfo/lingtyp</a><o:p></o:p></p>
                      </blockquote>
                    </div>
                    <p class="MsoNormal"><o:p> </o:p></p>
                  </div>
                </blockquote>
                <p class="MsoNormal"><br>
                  <br>
                  <o:p></o:p></p>
                <pre>-- <o:p></o:p></pre>
                <pre>Prof. Volker Gast<o:p></o:p></pre>
                <pre>English and American Studies<o:p></o:p></pre>
                <pre>Ernst-Abbe-PLatz 8<o:p></o:p></pre>
                <pre>D-07743 Jena<o:p></o:p></pre>
                <pre><o:p> </o:p></pre>
                <pre>Fon: ++49 3641 9-44546<o:p></o:p></pre>
                <pre>Fax: ++49 3641 9-44542<o:p></o:p></pre>
              </div>
              <p class="MsoNormal">_______________________________________________<br>
                Lingtyp mailing list<br>
                <a moz-do-not-send="true"
                  href="mailto:Lingtyp@listserv.linguistlist.org">Lingtyp@listserv.linguistlist.org</a><br>
                <a moz-do-not-send="true"
                  href="http://listserv.linguistlist.org/mailman/listinfo/lingtyp">http://listserv.linguistlist.org/mailman/listinfo/lingtyp</a><o:p></o:p></p>
            </div>
          </blockquote>
        </div>
        <p class="MsoNormal"><o:p> </o:p></p>
      </div>
      <br>
      <fieldset class="mimeAttachmentHeader"></fieldset>
      <br>
      <pre wrap="">_______________________________________________
Lingtyp mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Lingtyp@listserv.linguistlist.org">Lingtyp@listserv.linguistlist.org</a>
<a class="moz-txt-link-freetext" href="http://listserv.linguistlist.org/mailman/listinfo/lingtyp">http://listserv.linguistlist.org/mailman/listinfo/lingtyp</a>
</pre>
    </blockquote>
    <br>
    <pre class="moz-signature" cols="72">-- 
Martin Haspelmath (<a class="moz-txt-link-abbreviated" href="mailto:haspelmath@shh.mpg.de">haspelmath@shh.mpg.de</a>)
Max Planck Institute for the Science of Human History
Kahlaische Strasse 10   
D-07745 Jena  
&
Leipzig University 
IPF 141199
Nikolaistrasse 6-10
D-04109 Leipzig    





</pre>
  </body>
</html>