Dear All,
I'm sorry for writing when the heat of the discussion has been over 
for a week, but I'm just back after two weeks and having read the 
various contributions to the question I can't resist the temptation of 
providing an opinion myself.
As somebody has already pointed out this issue has been 
discussed many times already and for several decades . However, 
the diversity of approaches and opinions expressed as recently as 
on this present debate or at the already mentioned ICAZ session in 
Victoria proves that there is still space for much discussion and 
exchange of ideas. 
It is easy to say that all quantification systems are bad and 
therefore there is nor reason to debate them, but this is the nature 
of any archaeological evidence. It is fragmentary, incomplete and 
requires a bit of guess-work (also called interpretation).   
Unlike others I still see strong elements of disagreement between 
various zooarchaeologists, but at least we all seem to have 
digested the fact that the various quantification systems should  
not be taken as providing a solution but rather as tools that once 
compared with each other can help towards a solution. 
I find the question of MNI versus NISP still interesting because 
betrays the way different zooarchaeologists look at their material. 
As Keith Dobney and Paul Halstead have said, perhaps in different 
words, any counting of bones requires an element of decision 
making (what is a countable/identifiable  fragment?). Consequently 
the NISP ends up being an "interpretation of an interpretation" just 
as much as the MNI. 
Many bad things have been said about MNI, MNA and related 
methods, and I don't disagree with any, but how good or bad it is  
in comparison with NISP (or other systems) partly depends of 
whether we take into account what I regard as one of the main 
biases affecting the relative frequency of species: recovery. I find 
surprising that nobody has mentioned this factor. Virtually all times 
that I have had a chance to compare material from hand-collection 
and from sieving I find that the figures of MNI for the hand-collected 
assemblage approach those of the sieved assemblage much more 
than those for NISP. I am sure that everybody is aware of the fact  
that when cattle and sheep are compared the former is invariably 
better represented in the NISP count - undoubtedly an effect of a 
recovery bias. In this respect, since it contributes to reduce the 
effect of what I regard as the main bias in most assemblages I 
study,  I find the MNI a quite useful tool in getting an idea or the 
RELATIVE frequency of different taxa.
The main reason why I am mentioning this is not to advocate the 
use of the MNI, but just to provide an example of the fact no 
system provides raw, objective data that can easily be made 
available to everybody else. That we like it or not we all interpret 
and use subjective judgement. Therefore we could also go one step 
further and try to find ways to address some of the main biases 
that affect our data. 
Cheers,
Umberto
 





Umberto Albarella
Department of Ancient History and Archaeology
University of Birmingham
Edgbaston
Birmingham B15 2TT
U.K.
tel. +44/121/4147386
fax. +44/121/4145516
email [log in to unmask]
http://www.bham.ac.uk/BZL


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%