[Lingtyp] Kinship systems that distinguish age but not gender

Martin Haspelmath haspelmath at shh.mpg.de
Fri Jul 21 08:39:38 UTC 2017

It is indeed an interesting suggestion (by Bingfu Lu) that sex 
neutralization in kinship terms is related to the importance of sex for 
observers. This factor may also explain that we often have 
sex-differentiated terms for domestic animals, but rarely for wild animals.

But the "importance" of sex differentiation is not easy to assess. As 
Greenberg notes, there is a tendency to neutralize sex also in more 
remote relationships (e.g. with cousins, where even English neutralizes, 
and with in-laws), and it is hard to argue, for example, that sex is 
less important in cousins than in siblings. So maybe frequency of use is 
a better explanation after all? Does anyone have frequency counts for 
'younger sibling' and 'older sibling' terms? (And frequency counts for 
domestic as opposed to wild animals?)

I also have a comment on Maïa Ponsonnet's crictical remark concerning 
the term "universal":
> However, I wonder is calling such hypotheses "universals" too early 
> can create other problems. We may then omit to disqualify the 
> hypothesis, even after many, many counter-examples have been provided. 
> So we may end up postulating universality based on say, 10 cases, and 
> 10 years later still be busy providing counter-examples for what we 
> still call a "(potential) universal" while say, 20 counter-examples, 
> have already been provided.
> So perhaps calling it "hypothetical implication" may be safer?
The danger certainly exists that some claims become very famous and are 
repeated and believed even though there is no good evidence for them 
(e.g. that spinach contains a lot of iron).

But I feel that it is clear that every claim in science has the status 
of a hypothesis that is subject to potential disconfirmation. The 
differences reside in the amount of supporting evidence. The Konstanz 
Universals Archive <https://typo.uni-konstanz.de/archive/> is a great 
resource both for references to claims of universals and for the basis 
of the claims (thus, without reading Greenberg (1966), one can see that 
universal No 1656 is based on 120 languages).


On 21.07.17 01:16, bingfu Lu wrote:
> I agree with Martin’s bold claim.  It seems to be very natural in the 
> following senses.
> First, from the formal perspective, babies are very likely to be 
> neutralized in sex.  If there is a continuum of sex neutralization 
> from the point of being very young (babies) to the point of very old, 
> then, the younger section, which includes the babies, should be more 
> likely to be neutralized.
> Second,  from the perspective of linguistic iconicity, babies tend to 
> be sex-neutralized because their sex features are least developing. 
> And it is natural, the less sex-developing, the easier to be 
> sex-neutralized.
> According to the degrees of development in sex features, it might to 
> be predicted that there may be some languages where the very old 
> elders are neutralized in linguistic form, since very old elders are 
> sex-retrodegraded.
> In short, the sex neutralization is more likely when the sex features 
> are less strong and less important in age.
> ------------------------------------------------------------------------
> On Wednesday, July 19, 2017, 5:10:32 PM GMT+8, Martin Haspelmath 
> <haspelmath at shh.mpg.de> wrote:
> On the basis of Turkish (/kardeş/) and Minangkabau (/adiak/), which 
> neutralize the sex distinction in the younger sibling term, one could 
> propose the following universal:
> "If a language makes a distinction between elder and younger siblings 
> and neutralizes sex only in one type, then it neutralizes in younger 
> siblings."
> This may seem bold, but I think that such bold formulations are 
> productive in that they are likely to elicit responses from language 
> specialists whose language goes against the generalization. (And if 
> the bold generalization makes it into print somewhere, then one can 
> even write an abstract on the basis of one's data and argue against a 
> previous claim.)
> Now it so happens that a claim very similar to the one above has 
> already been made, on p. 76-77 in Greenberg's chapter "Universals of 
> kinship terminology", which is Chapter five of his most important work:
> Greenberg, Joseph H. 1966. /Language universals, with special 
> reference to feature hierarchies/. The Hague: Mouton.
> Greenberg formulates the generalization in terms of one kind of 
> kinship being "marked", the other "unmarked". "Marked" features tend 
> to be neutralized, so saying that younger siblings are "marked" 
> amounts to the same as the above claim. (In my view of things, this 
> would mean that some kinds of kinship features are more frequently 
> used than others.)
> (Greenberg also says somewhere that masculine/male is unmarked, so he 
> probably predicts that female terms ternd to be neuralized for age, 
> thus answering Siva Kalyan's question.)
> So there are a lot of interesting predictions that could be tested if 
> someone finally made a comprehensive world-wide database on kinship 
> terms (I think some people near Hedvig are working on this).
> Martin

Martin Haspelmath (haspelmath at shh.mpg.de)
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

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/lingtyp/attachments/20170721/364d6da6/attachment.htm>

More information about the Lingtyp mailing list