<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=Windows-1252">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<div class="moz-cite-prefix">Good question, Bernhard! I'm not sure what to say... Bybeean "relevance" is not easy to understand, and critics have long noted that it is a somewhat vague notion. It corresponds to the notion of a "functional sequence" in the "cartography"
and "nanosyntax" branches of generative grammar, but even the proponents of those approaches do not usually think that these sequences are part of an innate grammar blueprint.<br>
<br>
So it seems that this is about some kind of "close semantic relationship", and as you note, in my
<a moz-do-not-send="true" href="https://zenodo.org/record/1133888">2008 paper</a> ("Frequency vs. iconicity in explaining grammatical asymmetries"), I accepted explanations in terms of "iconicity of contiguity", e.g. of the relative ordering of aspect and tense
markers, or of number and case markers (where aspect/number are more closely related to the root meaning than tense/case, and also occur closer to the root).<br>
<br>
It is true that I rejected Haiman's "iconicity of cohesion", but I had good evidence only for the "cohesion contrasts" between his types (i) X-word-Y and (ii) X-Y; clearly, the presence of an additional marker is motivated by coding efficiency, not by "iconicity".
<br>
<br>
Now Haiman includes suppletion in this same scale (as the last item: (iii) Z), and this does not follow from efficiency of coding. Take coordination:<br>
<br>
(i) mother and aunt (X-word-Y)<br>
(ii) mother-father (X-Y)<br>
(iii) parents (Z)<br>
<br>
While the absence of "and" in (ii) may be explained by the predictability of the coordination relation due to frequency, the use of suppletion in (iii) seems to be due to *absolute* frequency (not relative frequency).<br>
<br>
It appears that we get suppletion only for "contiguous" meanings (this is explained by "locality" in generative grammar; see the
<a moz-do-not-send="true" href="http://www.crissp.be/bcgl-12-suppletion-allomorphy-and-syncretism/call-for-papers/">
upcoming Brussels conference on suppletion</a> etc.), and contiguity can be explained by iconicity.<br>
<br>
I'm not really sure about contiguity and compressing suppletion, but I still think that Haiman's notion of "cohesion" (for contrasts of the type (i) X-word-Y vs. (ii) X-Y) can be easily replaced by the well-understood notion of efficient coding.<br>
<br>
Best,<br>
Martin<br>
<br>
On 18.11.19 09:02, Bernhard Wälchli wrote:<br>
</div>
<blockquote type="cite" cite="mid:904ddea7088842af9d7dd6719f79aaea@ling.su.se"><style type="text/css" style="display:none;"><!-- P {margin-top:0;margin-bottom:0;} --></style>
<div id="divtagdefaultwrapper" dir="ltr" style="font-size:12pt;
color:rgb(0,0,0);
font-family:Calibri,Helvetica,sans-serif,EmojiFont,"Apple
Color Emoji","Segoe UI
Emoji",NotoColorEmoji,"Segoe UI
Symbol","Android Emoji",EmojiSymbols">
<div>Dear Martin,<br>
</div>
<br>
<div style="color:rgb(0,0,0)">
<div>
<div class="moz-cite-prefix">> Bybee (1985) showed that "more relevant" affix meanings are more likely to occur close to the root...</div>
<div class="moz-cite-prefix"><br>
</div>
<div class="moz-cite-prefix">
<div>How do you account for Bybee’s notion of “relevance”? By frequency? The way I read Bybee (1985) is that this at least originally was conceived of as iconicity: “If linguistic expression is iconic, then we would predict that the categories that are more
relevant to the verb will occur closer to the stem than those that are less relevant” (Bybee 1985: 24). I read this argument as follows: Since relevance is a semantic property, the iconicity assumption makes it testable empirically in form. </div>
<div>I cannot find any answer in Haspelmath (2008), where Bybee (1985) is mentioned just once (to strengthen a frequency argument) how Relevance relates to iconicity. Haspelmath (2008) is very useful in offering a typology of different approaches to iconicity.
Among those, there are two whose difference I never managed to understand: iconicity of cohesion: “Meanings that belong together more closely semantically are expressed by more cohesive forms” (this type is rejected in the paper; note also that its original
name given by Haiman was “conceptual distance”) and “iconicity of contiguity ("forms that belong together semantically occur next to each other; this is similar to iconicity of cohesion, but different in crucial ways”. This type is “beyond question” in Haspelmath
2008: 15; does “next to” mean "immediately adjacent" here, and that is why it is different from cohesion?). </div>
<div>Now, Bybee speaks of “closer” rather than “distance” (Haiman), but it seems to me that “more distant” is the antonym of “closer”, so am I wrong if I consider Bybee’s Relevance to be an instance of iconicity of conceptual distance? (cf. also “More specifically,
among the inflectional categories that we have surveyed, we would expect the most relevant to occur closest to the verb stem, and the least relevant to occur at the greatest distance from the verb stem”, Bybee 1985: 34-35).</div>
<div>Of course, it may be the case that the empirically tested hypothesis behind the notion of “relevance” still holds, but that the underlying assumption (“iconic”) was wrong. Relevance might then be an unexplained force like gravity. So, provided we assume
that Relevance is a fact, how is it explained?</div>
<div><br>
</div>
<div>Best wishes,</div>
<div><br>
</div>
<div>Bernhard W.</div>
<div><br>
</div>
<div><br>
</div>
<div>Bybee, Joan L. 1985. Morphology: A Study of the Relation between Meaning and Form. Amsterdam: Benjamins.</div>
<div>Haspelmath, Martin. 2008. Frequency vs. iconicity in explaining grammatical asymmetries." Cognitive linguistics 19.1: 1-33.</div>
<div><br>
</div>
<br>
</div>
</div>
</div>
</div>
</blockquote>
<br>
<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
Institut fuer Anglistik
IPF 141199
D-04081 Leipzig </pre>
</body>
</html>