[b-hebrew] We and us

David Kummerow farmerjoeblo at hotmail.com
Tue Nov 21 05:21:30 EST 2006

> On 21/11/2006 03:40, David Kummerow wrote:
> >/ ...
> />/
> />/ Now to you contention that "[l]ooking at other languages will only 
> />/ result in a lot of hot air". This is your view, and you are entitled to 
> />/ it. It is certainly not mine, and I would need convincing to move to 
> />/ such a position. Typology has demonstrated again and again that 
> />/ cross-linguistic variation in languages reveal systematic patterns ... I fail to see that this is "hot air"; rather, it is a 
> />/ concrete language predication. ...
> /
> Here I think we need to consider carefully what such patterns can teach 
> us. When we come to a new language, it is good to know what patterns to 
> expect. But we need to look at the data from that language in their own 
> right and find the language's own patterns. We should not come with the 
> presupposition that it will necessarily fit within the expected 
> patterns. If we do, we end up rejecting the data (such as Revell's) 
> simply because it doesn't fit our presuppositions. That is a very 
> dangerous scholarly procedure. It has been far too common in various 
> fields of scholarship for scholars to hypothesise a rule based on a 
> small number of examples, and then when collecting further examples, 
> some of which don't fit the hypothesis, to find some way to reject those 
> further examples. In this way faulty hypotheses rather too often find 
> themselves elevated to the status of scientific laws, at least until 
> someone comes along, re-evaluates the whole set of data, and finds that 
> the supposedly immutable rule was in fact based on almost nothing.

I agree strongly that each language must be considered on its own terms. 
Hebrew may prove to be an exception to the typological expectation; and 
that's fine. But it does allow for the formation of a hypothesis which 
can be tested on the data. I do not suggest the law is immutable (we are 
speaking of language, after all!), only that it is an expectation which 
can be tested.

> As an example of the danger of this in linguistic typology, consider the 
> one time firmly held typological rule that basic word order was never 
> object first. Then Des Derbyshire studied the Hixkaryana language of a 
> tiny tribe in the Amazon jungle and found data to suggest that its basic 
> word order was object first. Now if he had taken the approach you seem 
> to suggest, he would have concluded that this was in fact impossible and 
> so he must find some other way to explain his data. Fortunately he 
> didn't do this, but instead went on to study the phenomenon in detail, 
> to PhD level I think, and proved to the satisfaction of the linguistic 
> that, against the former typological rules, Hixkaryana is an object 
> first language. And in fact since then a few more object first languages 
> have been found - a tiny proportion of all the world's languages, but 
> sufficient to show that such typological rules should always be taken as 
> probabilities rather than certainties.

On word order, see Matthew Dryer's essays in the World Atlas of Language 
Structures. Hixkaryana certainly does not stand alone, even if 
object-first word order is cross-linguistically rare.

> >/ ... Further, it raises a helpful question 
> />/ concerning Revell's proposal and helps to raise the issues of 
> />/ evaluation: a) what is the nature of the synchronic politeness 
> />/ contrast(s); b) what is the diachronic development of the contrast(s), 
> />/ that is, can the development be traced to known politeness sources; and 
> />/ c) how has such a process occurred in BH when in other languages the 
> />/ process only happens after a distinction is made in the second person. 
> />/ Now, Revell's study has only touched in part on a). b) and c) are not 
> />/ discussed at all. But it is c) that is where it would get really 
> />/ interesting, for this would be getting at the heart of the function. But 
> />/ a study might flounder on completing a) such that a complete synchronic 
> />/ demonstration of a politeness distinction from the data might not be 
> />/ justifiable and hence b) and c) would be unnecessary.
> />/
> />/   
> /Of course it would be nice to be able to answer b) and c). But the 
> available data from Hebrew might be enough for us to demonstrate a), or 
> at least show that it is a probable explanation, without being able to 
> find out much about b) and c). I guess this is the position with 
> Hixkaryana, since it has no written record; we can simply observe that 
> it has this unusual word order without being able to find out how or 
> why. And the same could easily be true of Hebrew politeness distinctions.
> So, my real point here is that we should examine Revell's data 
> carefully, without presupposing on the basis of language typology that 
> it is necessarily wrong, or not to be explained in terms of politeness. 
> Then, when we have found out what is actually happening in Hebrew, it 
> would be good to compare it with Helmbrecht's rule, and consider whether 
> that rule needs to be modified in the light of new data.
> -- 
> Peter Kirk
> E-mail:  peter at qaya.org <http://lists.ibiblio.org/mailman/listinfo/b-hebrew>
> Blog:    http://speakertruth.blogspot.com/
> Website: http://www.qaya.org/

As I've said, it does not make Revell's analysis wrong since BH may be 
the exception to the typological prediction. But it does raise doubts 
such that another evaluation of the data is called for (Revell's study 
was incomplete in any case), and if possible points b) and c) should be 
given some discussion, as this would help to bolster a) if a politeness 
distinction was discovered. If the diachronic data does not exist, then 
b) is hard to work on, but it would need to be looked at. c) could be 
discussed in any case.

David Kummerow.

More information about the b-hebrew mailing list